The International Conference in Knowledge Management (CIKM'94), Gaithersburg, MD, November 1994.

Size: px
Start display at page:

Download "The International Conference in Knowledge Management (CIKM'94), Gaithersburg, MD, November 1994."

Transcription

1 The Intenational Confeence in Knowledge Management (CIKM'94), Gaithesbug, MD, Novembe 994. Hashing by Poximity to Pocess Duplicates in Spatial Databases Walid G. Aef Matsushita Infomation Technology Laboatoy Two Reseach Way Pinceton, New Jesey Hanan Samet Compute Science Depatment and Cente fo Automation Reseach and Institute fo Advanced Compute Studies The Univesity of Mayland College Pak, Mayland 2042 Abstact In a spatial database, an object may extend abitaily in space. As a esult, many spatial data stuctues (e.g., the quadtee, the cell tee, the R + -tee) epesent an object by patitioning it into multiple, yet simple, pieces, each of which is stoed sepaately inside the data stuctue. Many opeations on these data stuctues ae likely to poduce duplicate esults because of the multiplicity of object pieces. A novel appoach fo duplicate pocessing based on poximity of spatial objects is pesented. This is dieent fom conventional duplicate elimination in database systems because, with spatial databases, dieent pieces of the same object can span multiple buckets of the undelying data stuctue. xample algoithms ae pesented to pefom duplicate pocessing using poximity fo a quadtee epesentation of line segments and abitay ectangles. The complexity of the algoithms is seen to depend on a geometic classication of dieent instances of the spatial objects. By using poximity and the spatial popeties of the objects, the numbe of disk-i/o equests as well as the un-time stoage duing duplicate pocessing can be educed. Intoduction Spatial databases ae usually oganized in data stuctues that povide ecient access and exible manipulation of data. Thee ae seveal ways of epesenting a spatial object inside a data stuctue. Some data stuctues (e.g., the R-tee [9] and the Gid ile [2]) epesent a spatial object by just one entity inside the data stuctue (e.g., by the object's bounding ectangle in the case of the R-tee and by a point in a highe dimensional space in the case of the Gid ile). On the othe hand, anothe family of data stuctues (e.g., the quadtee [0], cell tee [8], and R + -tee [6]) epesent a spatial object by patitioning it into moe than one piece, each of which is stoed sepaately inside the data stuctue. In this pape, we focus on the latte family of data stuctues. We attempt to take advantage of the spatial chaacteistics of objects as a guide to duplicate pocessing (see [5] fo an oveview of techniques fo duplicate pocessing in database systems) that esult fom spatial database opeations. It is woth mentioning that duplicate pocessing in spatial databases is slightly dieent fom conventional duplicate elimination in database systems because with spatial databases, dieent pieces of the same object can span multiple buckets of the undelying data stuctue. o example, in the case of the R + -tee and the quadtee, duplicates aise because the object is patitioned into pieces. We can benet fom knowing that these pieces ae contiguous in space (if this is eally the case). In paticula, we dene an opeation, temed Repot Unique, which manages to pocess duplicates and epot each object in the data stuctue just once, egadless of the multiplicity of the patitions of the object inside the spatial data stuctue. Repot Unique can be viewed as an altenative appoach to hashing (see [] fo a good coveage of hashing techniques) when dealing with spatial data. Repot Unique maintains a dynamic data stuctue, temed the active bode [5], that seves as a epositoy fo the objects cuently being pocessed, temed active objects. The active bode data stuctue esembles a dynamic hash table. When conventional hashing techniques ae used thee is no way of pedicting when an object will not be efeenced again and hence can be safely deleted fom the hash table. Using spatial popeties of the undelying objects, the active bode can gow and shink in constant time. By knowing the extent and poximity of the objects, we can detect when all the pieces compising the object have been pocessed and hence can delete the object enty fom the active bode. This way we can avoid the situation that the hash table gows until it eaches its maximum limit, i.e., O(numbe of objects in the database). In fact, by using the active bode technique we can limit the stoage complexity to be in the ode of the active objects only. A look-up function, also based on spatial poximity, can be used to access enties in the active bode. The look-up function is the spatial analog of a hash function. As we demonstate in the pape, an impotant advantage of poximity-based look-up functions is that as a esult of thei use, buckets in the active bode ae guaanteed not to oveow. In addition, we also show howwe can utilize the poximity infomation to educe the numbe of disk-i/o equests duing duplicate pocessing. The est of the pape is oganized as follows. In Section 2 we descibe the data stuctues that we use fo stoing the undelying spatial database. The emaining sections descibe techniques to pocess duplicates using spatial

2 chaacteistics of the undelying spatial objects. Section descibes an algoithm that uses spatial poximity to eliminate the space equiements that esult fom using othe duplicate elimination techniques (e.g., hash tables). Section 4 demonstates that suppoting the Repot Unique opeation fo a ealistic vaiety of objects equies classifying spatial objects into categoies of dieent complexity. Section 5 shows how these classes aect the pefomance and coectness of the altenative appoaches to Repot Unique. Section 6 contains some concluding emaks. B s 2 The Undelying Spatial Data Stuctues We assume that the spatial objects ae stoed in data stuctues that patition the objects into multiple pieces. We use the quadtee as an example to demonstate ou algoithms fo duplicate pocessing. The quadtee patitions each object into multiple non-ovelapping pieces. Hee, we show how we oganize databases of line segments and ectangles using vaiants of the quadtee. o line segments we use the PMR quadtee [], while fo ectangles we useavaiant of the egion quadtee [0], temed the ectangle quadtee. In the PMR quadtee, a line segment is divided into seveal pieces, whee each piece is associated with the quadtee block that it intesects. igue shows one example of the PMR quadtee. A block in the PMR quadtee is pemitted g d h i igue : An example of a PMR quadtee. to contain a vaiable numbe of line segments. It is constucted by inseting the line segments one-by-one into an initially empty stuctue consisting of one block. ach line segment is inseted into all of the blocks that it intesects o occupies in its entiety. Duing this pocess, the occupancy of each aected block ischecked to see if the insetion caused it to exceed a pedetemined splitting theshold. If the splitting theshold is exceeded, then the block is split once, and only once, into fou blocks of equal size. ach block in the PMR quadtee stoes the object identies of the lines passing though it. The full desciption of the line segments (e.g., the stat and end coodinate values of the end-points) is stoed in what is called the featue table. The index of line l in the featue table is l's object identie. Notice that we cannot simply tavese the featue table and epot the lines that we nd since not all the lines in the featue table have to belong to the data stuctue in question, i.e., that the featue table may be shaed by moe than one PMR quadtee. Rectangles ae stoed in a vaiant of the egion quadtee [0] which we tem a ectangle quadtee. ach ectangle is decomposed into the quadtee blocks that lie inside the ectangle. o example, igue 2a shows the decomposition of ectangle into its constituent quadtee blocks. We assume that the ectangles need not be disjoint. We also adopt the estiction that each quadtee block is completely inside all of the ectangles that ovelap it. In othe wods, the case a c b e f igue 2: The decomposition of ectangle into its constituent quadtee blocks. Block B is decomposed afte ectangle s is inseted. that only pat of a quadtee block lies in a ectangle is not allowed. o example, this estiction means that block B of igue 2a must be decomposed when ectangle s is added to yield the decomposition given in igue 2b. The ectangle quadtee epesentation assumes that all the blocks that ae intenal to a ectangle stoe an object identie which isan oset in a featue table whee the uppe-left and bottomight coodinate values of the ectangle ae stoed. A cental opeation in spatial databases is the window opeation. A window can be in the fom of a ectangle o a polygon. This opeation seves as a building block fo anumbe of queies. Usually, spatial featues span a wide featue space. Howeve, uses ae moe inteested in viewing o queying only potions of the featue space instead of the whole space. xtacting pats of the space to wok with in subsequent opeations is done by windowing. Given a window w, some examples of window-based queies ae: epot all featues inside w, intesect featue f with featue b only inside w, detemine if featue f exists in w, etc. Moe details about window queies and algoithms fo answeing them can be found in []. When a ectangle is clipped (e.g., as a esult of a window opeation), the ectangle quadtee does not update the coodinate values of the the clipped ectangle in the featue table. Notice that we cannot simply tavese the featue table and epot the ectangles that we nd since not all the ectangles in the featue table necessaily belong to the data stuctue in question. Avoiding the xta Space - A ist Attempt In this section we pesent a simple duplicate pocessing algoithm that does not equie additional space (e.g., in contast to the O(n) space equied fo hashing in the case of a database of n objects). We also discuss the limitations of this algoithm. In the emaining sections of the pape, we will addess these limitations and ty to tackle them. We pesent the algoithm fo the case of a database of line segment objects. A simila algoithm can be descibed fo a database of ectangula objects. In ode to exploit the poximity of object pieces we de- ne a test function, say t, such that given a spatial object, say o, with k pieces, p, p 2,..., p k, t hasavalue of false fo only one of the pieces, say p j of o. Application of t to the est of the pieces yields the value of tue. Any function t that satises this citeion can be used as a means of suppessing all pieces of o othe than p j fom epoting o's object identie, and hence can be adopted by the Repot Unique opeation. o example, in the case of hashing,

3 a function t is dened as follows (a is a bit aay with all its elements initialized to have avalue of false, and oid(p i) is the identie of the object to which piece p i belongs): t (p i) = et value a[oid(p i)]; if(not a[oid(p i)]) then a[oid(p i)] tue; etun (et value); We can get id of the O(n) space equiements of the function t in the following way. This gives ise to algoithm Repot Unique descibed below. We tavese the blocks of the PMR quadtee, and fo each block, say B, we pefom the following actions:. o each object identie i stoed in B (that coesponds to piece p i), etieve the line segment desciption, say l i, fom the featue table, and 2. pefom the following testing function t 2 t 2(l i;b) = if(stat point(l i) in B) then etun (false) else etun (tue). If t 2(l i;b) is false, then epot line segment l i. We use igue fo illustation. Notice that fo each line segment l, t 2 is false only when the block containing the stat point oflis pocessed. Application of t 2 to all blocks containing pieces of othe than the uppe-left cone of yields tue (t 2 is efeed to as the point-in-block test). Notice also that the algoithm woks only in the case whee the pieces compising the spatial object ae contiguous. The case whee the pieces of an object ae disjoint (e.g., due to clipping pats of the object by a window opeation, descibed in Section 2) is handled in Section 5. p B l q B 2 B igue : Line l is stoed in blocks B, B 2, and B whee only block B causes the point-in-block test to succeed since the stating point p of l only lies inside B and not in B 2 no in B. The algoithm Repot Unique does not equie any additional space. Howeve, its CPU and disk-i/o costs ae high. The algoithm applies a point-in-block test pe object piece, which equies fou compaisons (each of cost C cpu). In paticula, the point (x; y) is inside block B((bl x;bl y), (u x;u y)), whee bl and u denote the bottom left and uppe ight cones of the block, espectively, i bl x x u x ^ bl y y u y. t 2 is pefomed nk times, whee n is the numbe of objects in the database, and k is the aveage numbe of pieces pe object in the database. Notice that in ode to pefom the point-in-block test (t 2)wemust access the featue table (via the object identi- e) each time we encounte a piece of a line segment. This is necessay in ode to etieve the stat point of the line segment. Assuming that the cost of etieving a segment fom the featue table is C io, then the execution time of the algoithm is 4knC cpu + knc io. The cost tem knc io involves edundant disk-i/o equests. Basically, it epesents the cost of etieving line segments fom the featue table (once fo evey line patition, amounting to k disk equests pe line segment). A bette algoithm would pefom only n such equests, i.e., one equest pe line segment in the database instead of one equest fo each piece of each line segment. Thus, we would like to develop an algoithm that does not need the O(n) exta stoage (o at least having an asymptotic stoage cost less than O(n)), yet one that still pefoms only n disk equests (e.g., n accesses to the featue table). In the following sections, we show howwe can achieve this goal. Howeve, this depends on a close scutiny of the natue of the spatial objects which is the subject of the next section. 4 Object Classication An impotant facto aecting the pefomance of the Repot Unique opeation is the natue of the objects stoed in the database. o example, what is the eect of esticting the lines to be ectilinea, in contast to lines with abitay slopes? As anothe example, suppose that objects ae patially clipped as is the situation afte a window opeation (descibed in Section 2). Is it moe dicult to epot these clipped objects than epoting non-clipped objects? As an illustation of the second example, note that due to the way the PMR quadtee is dened, Repot Unique does not wok popely if the line segments ae patially clipped. This is shown in in igue 4. In paticula, when a line is clipped (e.g., as a esult of a window opeation), the PMR quadtee does not update the stating and ending points of the clipped line in the featue table. This is done since the line may be shaed by othe indexes in the database. In addition, this ensues consistency when potions of line segments ae emoved and then ae added back late in the pocessing as a esult of set opeations [6]. Thus the featue table is not updated and the clipped line segment is stoed implicitly. As a esult, if a line, say l, in igue 4 is clipped so that only s emains and if s does not contain l's stating point, then s will not be epoted by Repot Unique. This case suggests that we have to conside the altenative classes of objects (e.g., clipped objects) when developing algoithms fo Repot Unique since it aects the coectness of the algoithm. Due to space limitations, we estict ou discussion to ectangula objects only. Moe details about the classication and duplicate pocessing algoithms fo line segment objects can be found in [2]. Spatial objects of type ectangle can be classied in the following way (the dieent classes ae given in igue 5): Class- ectangles: a ectangle having no clipped o missing potions as illustated in igue 5a, and temed a egula ectangle. Class-2 ectangles: a ectangle whee only a ectangula potion of it is included and the est of the ectangle is clipped as illustated in igue 5b, and temed a clipped ectangle. Class- ectangles: a ectangle whee seveal potions (holes) may be missing, but still the est of the ectangle is connected, as illustated in igue 5c, and temed a clipped-out ectangle. Class-4 ectangles: a ectangle whee seveal potions (holes) of abitay shape may be missing that esult

4 q p x p y q x q y q p x p y q x q y p l p l B B 2 B B B 2 B featue table clipping window featue table igue 4: Line l as stoed in a PMR quadtee. The esult of clipping l by a window opeation in a PMR quadtee. The oiginal stating point oflin the featue table is neithe in blocks B 2 no in B. Notice that the featue table still stoes point p as pat of the desciption of the clipped line. in the ectangle being decomposed into seveal disjoint pieces, as illustated in igue 5d, and temed a disjoint ectangle. Although disjoint, all the potions of the ectangle efe to, and epesent, just one object of type ectangle. Type- ectangles: ectilinea ectangles, i.e., whose sides ae paallel to the axes (igue a). Type-2 ectangles: abitay ectangles - i.e., ectangles whose othogonal sides have abitay oientations (slopes) (igue b). (c) (d) igue 5: Classication of ectangles Class-, Class- 2, (c) Class-, (d) Class-4. This classication of ectangles is of pactical use. Class- epesents egula ectangles (e.g., bounding boxes of some spatial objects in a map). Class-2 epesents the pats of a ectangle that emain afte a ectangula window opeation (igue 6a). Class- epesents the pats of a ectangle window window window Clipped out egions window clipping polygon (c) clipping polygon igue 6: xample illustating how some of the ectangle classes ae ceated: Class-2, Class-, (c) Class-4. that lie outside one o moe ectangula quey windows (igue 6b). These ectangles esult fom clipping-out egula ectangles against the quey ectangles. Notice that each of the esulting ectangles is still epesentable as a fouconnected egion. Class-4 epesents the pats of a ectangle that lie inside (o outside) one o moe quey windows of abitay shape. o example, igue 6c is the esult of intesecting and clipping a ectangle with a numbe of simple polygons. We futhe classify ectangles of each class into the following types accoding to thei oientation: igue : Two types of ectangles: Type-, Type-2. 5 Duplicate Pocessing fo Rectangula Objects Below, we pesent algoithms fo the Repot Unique poblem that can handle seveal class/type combinations of ectangula objects given in Section 4, and see how these combinations aect the complexity of the algoithms. It is impotant to obseve that ectangles of the undelying database ae allowed to ovelap in space. Let q be the maximum numbe of ovelapping ectangles at any point in space, i.e., fo any block, thee ae at most q ovelapping ectangles. 5. Class- Type- and Class-2 Type- Rectangles The ectangles that coespond to these class/type combinations ae ones whose sides ae paallel to the x and y axes whee some of the ectangles may be clipped (Class-2) but ae still of ectangula fom. Ou goal fo this Class/Type combination is not to pefom any disk I/O to the featue table. In this case, the only infomation at hand ae the ectangle identies in the quadtee blocks. The algoithm taveses the ectangle quadtee blockby-block and maintains the active set of ectangle identi- es. These coespond to the ectangles that intesect the block. A ectangle identie, say id, which coesponds to a ectangle, say, is added to the active set when id is st encounteed duing the tavesal. id is deleted fom the active set once all the quadtee blocks that intesects have been visited. The necessay stoage is bounded by the maximum size of the active set duing the execution of the algoithm. The size of the active set is consideably smalle than n, the numbe of ectangles in the spatial database.

5 In ode to execute the algoithm eciently, we need to oganize the active set of ectangles. The oganization of the active set depends on the opeations that ae to be pefomed on the set. The algoithm pefoms the following thee pimitive opeations on an active set, say A: () test if a ectangle identie exists in A (notice that this is needed in ode to decide whethe a ectangle has aleady been encounteed o if it is encounteed fo the st time in the tavesal), (2) inset a ectangle identie into A, and () delete a ectangle identie fom A. Befoe poviding moe details about the algoithm, obseve that if we ae able to pefom each of the above thee opeations in O() time and maintain the undelying data stuctue that stoes the active set in O() time, then the oveall complexity of the algoithm would be O(nk) time with O(active set size) space. The algoithm visits the blocks of the quadtee in NW, N, SW, S scanning ode. It maintains one basic data stuctue: an active bode [5]. In the context of duplicate pocessing and hashing, the active bode is the spatial analog of a hash table. The active bode epesents the bode between those quadtee blocks that have been pocessed and those that have not. The elements of the active bode fom a \staicase" of vetical and hoizontal edges, moving fom southwest to notheast, as shown by the heavy line in igue 8. ach element of the active bode coesponds to an edge (bode) in the active bode. Initially, the active bode consists of two elements that coespond to the noth and west bodes of the image. When the algoithm teminates, the active bode consists of elements that coespond to the south and east bodes of the entie image. In othe wods, wheneve a node is visited by the algoithm, the elements of the active bode ae updated accodingly to include the bode of this new block. We dene an active ectangle as a ectangle that is patially pocessed, i.e., at least one of the quadtee blocks ovelapping with this ectangle has not been pocessed yet. A ectangle is inactive if eithe all o none of the blocks though which it passes have been pocessed by the tavesal algoithm. Thee is a data bucket associated with each element of the active bode. ach bucket is of capacity q, which is the same as the bucket capacity of the undelying quadtee block. vey element of the active bode stoes in its bucket the set of ectangle identies which coespond to the active ectangles that intesect the potion of the active bode coesponding to this element. Rectangle identies ae added to the active bode when they ae encounteed duing the tavesal, and ae popagated into the active bode until all the pieces of a ectangle have been visited. A ectangle is epoted when it is being deleted fom the active bode. The poblem is how tode- tect that a ectangle has been pocessed in its entiety when thee is no means fo the algoithm to know the coodinate values of the end-points of the ectangle. In ode to achieve this, the algoithm maintains some tempoay infomation in the active bode to help it detemine that the lowe-ight cone of the ectangle has been eached and that the ectangle identie can be epoted. o example, conside the ectangle given in igue 2a. We use igue 8 to illustate the pocess of epoting its pesence. Heavy shading is used to indicate that the block has aleady been tavesed. When block A is encounteed duing the tavesal (igue 8a), 's identie (associated with A) is inseted into the active bode. When block B is visited (igue 8b), 's identie is popagated into the active bode. Upon encounteing a westen o southen bounday of a ectangle fo the st time (e.g., when pocessing block C in igue 8c) a special make symbol, say e, is associated with the active bode element that led to the detection of this fact (e.g., the active bode element that is adjacent to block C). Notice that e is popagated into the active bode when block D is pocessed (igue 8d). As anothe example, we obseve that a special make symbol, say s is added to the active bode element contiguous to block (igue 8e) and is popagated into the active bode elements to the east of, when thei adjacent blocks in the database ae pocessed. At the time block is pocessed (igue 8f), the active bode elements to the noth and west of will contain the special makes e and s, espectively. This situation seves to indicate that we ae though. At this point, both e and s ae deleted fom the active bode and 's identie is epoted. Notice that is epoted once all the blocks compising have been visited by the tavesal pocedue. o example, in igue 8f, at the time block is visited, we ae sue that all the blocks inside ae aleady pocessed by the algoithm. This is because a NW, N, SW, S tavesal ode is admissible [4]. The complexity of the algoithm can be evaluated as follows. When a block in the quadtee that coesponds to a leaf node is pocessed, the potion of the active bode that is adjacent to the block must be located. In [], a technique is descibed that enables the blocks to be located in the active bode in constant time. This is achieved by tavesing (and updating) the active bode along with the quadtee tavesal while passing and stacking pointes to guaantee that the exact location in the active bode is available (O() time) wheneve needed, so that seaching in the active bode is entiely avoided. The eade is efeed to [] fo futhe details. Once the appopiate active bode element islo- cated, testing this element fo the existence of a ectangle as well as insetion and deletion of a ectangle can each be done in O(q) time which is eally a constant oo(). As demonstated in this section, ou algoithm does not need to pefom any disk I/O equests to access the featue table. As a esult, the time complexity of this algoithm is O(nkq) which is popotional to the numbe of object pieces in the database that ae encounteed duing the tavesal. Notice that the facto q is included because all the pocedues that access the active bode pefom in O(q) time, as descibed in the intoduction of Section 5. The space complexity iso(active ectangles), whee the active ectangles ae the ones that intesect the active bode at any given point. The size of the active set in the wost case is O(qT) whee q is the maximum numbe of objects that can be stoed in a quadtee block befoe it oveows (i.e., the bucket size) and T is the width (e.g., pixels) of the space compising the undelying spatial database. In pactice, it is expected that the size of the active set is consideably smalle than O(qT). Obseve that it is guaanteed that the active bode buckets will neve oveow since thee is one-to-one coespondence between elements of the active bode and blocks of the undelying quadtee. Since each block can hold up to q ectangles, then at wost each bucket will hold the same numbe of ectangles and hence cannot oveow. Notice that the algoithm does not depend on the actual coodinate values of the ectangle since it does not access the enties in the featue table and hence it woks coectly even if ectangles ae clipped due to some intesection with a ectangula window (Class-2 ectangles).

6 active bode active bode active bode A B C D A B C D A B Ce D (c) active bode active bode active bode A B Ce De A B C De A B C D s e e s s (d) (e) (f) = visited block = block of the ectangle igue 8: When block A is pocessed is inseted into the active bode. is popagated into the active bode when block B is pocessed. (c) When block C is pocessed a special make symbol e is inseted into the active bode. (d) e is popagated into the active bode when block D is pocessed. (e) When block is pocessed a special make symbol s is inseted into the active bode. (f) is epoted at the time block is pocessed and both e and s ae deleted fom the active bode. 5.2 Class- Type- and Class-4 Type- Rectangles Class- and Class-4 imply that the ectangles may have notches on them o may even consist of disjoint pieces. o these classes, we cannot avoid accessing thei enties in the featue table in ode to know the eal extent of the ectangles while tavesing them. Rectangle identies ae popagated in the active bode (egadless of whethe we ae tavesing a pat of the ectangle o a clipped out potion of the ectangle) until the lowe-ight cone of the ectangle is visited by the tavesal. In this case, the ectangle is epoted and deleted fom the active bode. This implies that the algoithm has to issue some disk I/O equests to the featue table. The algoithm poceeds as follows:. Tavese the spatial database as well as the active bode simultaneously in the NW, N, SW, S ode. 2. Maintain in the active bode: the identies of the ectangles that intesect the active bode o whose pieces have not yet been pocessed in thei entiety by the algoithm, and the coodinate values of the end-points of the ectangles in.. When the block containing the lowe-ight cone of ectangle is eached, epot the identie of, and delete 's infomation fom the active bode. igue 9 illustates the locations at which a disk I/O equest is issued and when a ectangle is epoted fo seveal possible ectangles in Class- and Class-4. Notice that fo Class- ectangles one, o moe disk I/O equests can be issued. o example igues 9a-c and e have only one disk I/O equest, igue 9d has two disk I/O equests as the uppe-left cone of the ectangle is clipped-out, while igue 9f has many disk I/O equests. o Class-4 ectangles, moe than one disk I/O equest is possible (e.g., igue 9g) depending on the numbe of blocks in one of the fou line segments of the window bounday that has a N-SW oientation and is closest to the oigin of the undelying space. Notice that fo Class-4 ectangles, if the uppe-left cone of the ectangle is not clipped out, then only one disk I/O equest is issued egadless of the numbe of disjoint pieces of the ectangle (e.g., igue 9h). Since the uppe-left cone of the ectangle is pocessed st, a disk I/O equest is issued to the featue table and the coodinate values of the end-points of the ectangle ae made known to the active bode. This helps avoid any futhe disk I/O equests. Obseve that this is not the case when the uppe-left cone of the ectangle is clipped out (e.g., igue 9f). igues 9f and 9g illustate the cases whee the maximum numbe of disk I/O equests fo Class- and Class-4 ectangles can be geneated, espectively. Let k w be the numbe of blocks in the edge of the window with a N{SW oientation that is closest to the oigin of the undelying space. igue 0 illustates the wost case numbe of disk I/O equests that ae equied with lines of this oientation. Obseve that a disk I/O equest occus with evey othe piece in the object. Theefoe, d kw e disk I/O equests ae 2 issued. Then, the maximum numbe of disk I/O equests to the featue table fo Class- and Class-4 is d kw e. 2

7 (c) (d) = disk I/O equest igue 9: -(c) When the ectangle is st encounteed a disk I/O equest is issued to etieve the ectangle's coodinate values. (d) Two disk equests ae issued wheneve, the uppe-left cone of the ectangle is clipped out (Class- ). (e) Only one disk equest is issued since the uppe-left cone of the ectangle is not clipped out (Class-). (f) The wost-case fo Class- ectangles occus when many quey windows clip out pats of the ectangle such that the uppeleft piece of the emaining potion of the ectangle foms a line segment of N-SW oientation. (g) Redundant disk equests ae issued when the uppe-left piece of the ectangle is pocessed since the uppe-left cone of the ectangle is clipped out (Class-4). (h) Only one disk equest is issued although the ectangle is patitioned into moe than one disjoint piece (Class-4). Notice that the uppe-left cone of the ectangle is not clipped out. It is woth mentioning that the task of uniquely epoting spatial objects is dieent fom that of connected component labeling [, 4]. In the latte task we ae inteested in assigning a unique identie to each fou- (o eight-) connected egion in the undelying space. The pincipal dieence is that, geneally speaking, the inputs of the two tasks ae diffeent. Moe specically, in Repot Unique we can always pefom a disk I/O equest to the featue table and etieve the object's exact desciption. This infomation is not made available to the connected component labeling algoithms. In fact, the main pupose of the connected component labeling algoithms is to build the object's exact desciption fom the local infomation given in the undelying space. As a esult of the dieence in the natue of the type of input and the goals of the two tasks, thei complexities ae (e) = disk I/O equest igue 0: The maximum numbe of disk equests that can aise fo a line with N-SW oientation. (f) (g) (h) dieent. o example, igue demonstates the wostcase complexity fo the connected component labeling algoithm (measued in tems of the new labels that have to be intoduced). On the othe hand, the shape of the e- B C 6 A igue : An 8 8 image that esults in geneating a maximum numbe of equivalence pais fo the connected component labeling algoithm. gion in igue does not epesent the wost case disk I/O complexity fo Repot Unique. In paticula, only two disk I/O equests will be issued (i.e., when blocks A and B ae visited). The wost case numbe of disk I/O equests fo Repot Unique fo clipped-out ectangles (Class-) is given in igue 9f. Notice that having a saw-tooth-like shape (e.g., as in igue ) does not change the wost case since at block B of igue, a disk I/O equest is issued to etieve the exact desciption of the ectangle fom the featue table, and this desciption will be popagated in the active bode in the easten and southen diections. This coves all the sawtooths to the east and south of block B, theeby obviating the need fo additional disk I/O equests (e.g., at block C). 5. Rectangles of Type-2 In ode to ensue that ectangles of abitay oientations ae epoted only once, we maintain thei ectilinea bounding box in the active bode (igue 2a) and epot the ectangle identie once the lowe-ight cone of the bounding box is coveed by the tavesal. As descibed in Section 5.2 = disk I/O equest igue 2: A bounding box is stoed in the active bode fo ectangles of Type-2. Seveal disk equests occu when pocessing the line segment closest to the uppe-left cone. (also in igues 9f, 9g, and 0) the numbe of disk I/O equests to the featue table is popotional to d kw e, whee 2 k w is the numbe of blocks in the line segment of the window bounday with a N-SW oientation that is closest to the oigin of the undelying space (igue 2b). Notice that once this set of disk I/O equests is issued, the algoithm does not need to access the featue table again since the

8 popagation of the ectangle identie and coodinate values in the active bode will help avoid any futhe disk I/O equests to the featue table. 6 Conclusion and utue Reseach Table 6 gives a summay of ou esults. It shows the dieent complexities of the algoithms developed fo the vaious class/type object combinations. As we can see, duplicate Type- Type-2 Class- O(knq) avg. cpu, no i/o O(knq) avg. cpu, d kw 2, en i/o Class-2 O(knq) avg. cpu, no i/o O(knq) avg. cpu, d kw 2, en i/o Class- O(knq) avg. cpu, d kw 2 en i/o O(knq) avg. cpu, d kw 2 en i/o Class-4 O(knq) avg. cpu, d kw 2 en i/o O(knq) avg. cpu, d kw 2 Table : en i/o Summay of time and space complexities of Repot Unique fo class/type combinations of ectangle objects. pocessing using the spatial chaacteistics of objects is a challenging poblem. In paticula, using the spatial chaacteistics of the objects enables us to adapt techniques used with hashing. o example, we wee able to educe the numbe of disk I/O equests (e.g., in compaison to O(nk) disk I/O equests when using hashing), sometimes to 0 o O(n) which is a signicant impovement. In addition, by using the extent and poximity of spatial objects we wee able to detect when an object is no longe needed in the active bode (analogous to a hash table), and hence we wee able to bound the size of the table. This was achieved by deleting the objects that ae entiely pocessed by the algoithm. utue wok involves developing bette algoithms fo othe class/type combinations, classifying spatial objects besides line segments and ectangles as well as developing algoithms fo them, and consideing othe spatial data stuctues that patition objects besides the quadtee vaiants. Acknowledgements The st autho conducted most of this eseach while at The Univesity of Mayland, College Pak. The suppot of the National Science oundation unde Gant IRI is gatefully acknowledged. Refeences [] W. G. Aef and H. Samet. cient pocessing of window queies in the pyamid data stuctue. In Poceedings of the 9th. ACM SIGACT-SIGMOD- SIGART Symposium on Pinciples of Database Systems (PODS), pages 265{22, Nashville, TN, Apil 990. (also in Poceedings of the ifth Bazilian Symposium on Databases, Rio de Janeio, Bazil, Apil 990, 5{ 26). [2] W. G. Aef and H. Samet. Uniquely epoting spatial objects: Yet anothe opeation fo compaing spatial data stuctues. In Poceedings of the 5th Intenational Symposium on Spatial Data Handling, pages 8{89, Chaleston, SC, August 992. [] M. B. Dillencout and H. Samet. Using topological sweep to extact the boundaies of egions in maps epesented by egion quadtees. Submitted fo publication, 99. [4] M.B. Dillencout, H. Samet, and M. Tamminen. A geneal appoach to connected-component labeling fo abitay image epesentations. Jounal of the ACM, 9(2):25{280, Apil 992. [5] R. lmasi and S. B. Navathe. undamentals of Database Systems. Benjamin/Cummings, Redwood City, CA, 989. [6] C. aloutsos, T. Sellis, and N. Roussopoulos. Analysis of object oiented spatial access methods. In Poceedings of the 98 ACM SIGMOD Intenational Confeence on Management of Data, pages 426{49, San ancisco, May 98. [] Michael J. olk and Bill Zoellick. ile Stuctues - Second dition. Addison-Wesley, Reading, MA, 992. [8] O. Gunthe. The design of the cell tee: an objectoiented index stuctue fo geometic databases. In Poceedings of the ifth I Intenational Confeence on Data ngineeing, pages 598{605, Los Angeles, ebuay 989. [9] A. Guttman. R-tees: A dynamic index stuctue fo spatial seaching. In Poceedings of the 984 ACM SIGMOD Intenational Confeence on Management of Data, pages 4{5, Boston, June 984. [0] A Klinge. Pattens and seach statistics. In J. S. Rustagi, edito, Optimizing Methods in Statistics, pages 0{. Academic Pess, New Yok, 9. [] R. C. Nelson and H. Samet. A consistent hieachical epesentation fo vecto data. Compute Gaphics, 20(4):9{206, August 986. (also Poceedings of the SIGGRAPH'86 Confeence, Dallas, August 986). [2] H. Nievegelt, H. Hintebege, and K. C. Sevcik. The gid le: an adaptable, symmetic multikey le stuctue. ACM Tansactions on Database Systems, 9():8{, Mach 984. [] A. Rosenfeld and A. C. Kak. Digital Pictue Pocessing. Academic Pess, New Yok, second edition, 982. [4] H. Samet. Connected component labeling using quadtees. Jounal of the ACM, 28():48{50, July 98. [5] H. Samet. Applications of Spatial Data Stuctues: Compute Gaphics, Image Pocessing, and GIS. Addison-Wesley, Reading, MA, 990. [6] H. Samet. The Design and Analysis of Spatial Data Stuctues. Addison-Wesley, Reading, MA, 990.

Journal of World s Electrical Engineering and Technology J. World. Elect. Eng. Tech. 1(1): 12-16, 2012

Journal of World s Electrical Engineering and Technology J. World. Elect. Eng. Tech. 1(1): 12-16, 2012 2011, Scienceline Publication www.science-line.com Jounal of Wold s Electical Engineeing and Technology J. Wold. Elect. Eng. Tech. 1(1): 12-16, 2012 JWEET An Efficient Algoithm fo Lip Segmentation in Colo

More information

Shortest Paths for a Two-Robot Rendez-Vous

Shortest Paths for a Two-Robot Rendez-Vous Shotest Paths fo a Two-Robot Rendez-Vous Eik L Wyntes Joseph S B Mitchell y Abstact In this pape, we conside an optimal motion planning poblem fo a pai of point obots in a plana envionment with polygonal

More information

Controlled Information Maximization for SOM Knowledge Induced Learning

Controlled Information Maximization for SOM Knowledge Induced Learning 3 Int'l Conf. Atificial Intelligence ICAI'5 Contolled Infomation Maximization fo SOM Knowledge Induced Leaning Ryotao Kamimua IT Education Cente and Gaduate School of Science and Technology, Tokai Univeisity

More information

DYNAMIC STORAGE ALLOCATION. Hanan Samet

DYNAMIC STORAGE ALLOCATION. Hanan Samet ds0 DYNAMIC STORAGE ALLOCATION Hanan Samet Compute Science Depatment and Cente fo Automation Reseach and Institute fo Advanced Compute Studies Univesity of Mayland College Pak, Mayland 07 e-mail: hjs@umiacs.umd.edu

More information

RANDOM IRREGULAR BLOCK-HIERARCHICAL NETWORKS: ALGORITHMS FOR COMPUTATION OF MAIN PROPERTIES

RANDOM IRREGULAR BLOCK-HIERARCHICAL NETWORKS: ALGORITHMS FOR COMPUTATION OF MAIN PROPERTIES RANDOM IRREGULAR BLOCK-HIERARCHICAL NETWORKS: ALGORITHMS FOR COMPUTATION OF MAIN PROPERTIES Svetlana Avetisyan Mikayel Samvelyan* Matun Kaapetyan Yeevan State Univesity Abstact In this pape, the class

More information

A Memory Efficient Array Architecture for Real-Time Motion Estimation

A Memory Efficient Array Architecture for Real-Time Motion Estimation A Memoy Efficient Aay Achitectue fo Real-Time Motion Estimation Vasily G. Moshnyaga and Keikichi Tamau Depatment of Electonics & Communication, Kyoto Univesity Sakyo-ku, Yoshida-Honmachi, Kyoto 66-1, JAPAN

More information

SORTING IN SPACE HANAN SAMET

SORTING IN SPACE HANAN SAMET SORTING IN SPACE HANAN SAMET COMPUTER SCIENCE DEPARTMENT AND CENTER FOR AUTOMATION RESEARCH AND INSTITUTE FOR ADVANCED COMPUTER STUDIES UNIVERSITY OF MARYLAND COLLEGE PARK, MARYLAND 074-4 USA e mail: hjs@cs.umd.edu

More information

ART GALLERIES WITH INTERIOR WALLS. March 1998

ART GALLERIES WITH INTERIOR WALLS. March 1998 ART GALLERIES WITH INTERIOR WALLS Andé Kündgen Mach 1998 Abstact. Conside an at galley fomed by a polygon on n vetices with m pais of vetices joined by inteio diagonals, the inteio walls. Each inteio wall

More information

An Unsupervised Segmentation Framework For Texture Image Queries

An Unsupervised Segmentation Framework For Texture Image Queries An Unsupevised Segmentation Famewok Fo Textue Image Queies Shu-Ching Chen Distibuted Multimedia Infomation System Laboatoy School of Compute Science Floida Intenational Univesity Miami, FL 33199, USA chens@cs.fiu.edu

More information

Input Layer f = 2 f = 0 f = f = 3 1,16 1,1 1,2 1,3 2, ,2 3,3 3,16. f = 1. f = Output Layer

Input Layer f = 2 f = 0 f = f = 3 1,16 1,1 1,2 1,3 2, ,2 3,3 3,16. f = 1. f = Output Layer Using the Gow-And-Pune Netwok to Solve Poblems of Lage Dimensionality B.J. Biedis and T.D. Gedeon School of Compute Science & Engineeing The Univesity of New South Wales Sydney NSW 2052 AUSTRALIA bbiedis@cse.unsw.edu.au

More information

a Not yet implemented in current version SPARK: Research Kit Pointer Analysis Parameters Soot Pointer analysis. Objectives

a Not yet implemented in current version SPARK: Research Kit Pointer Analysis Parameters Soot Pointer analysis. Objectives SPARK: Soot Reseach Kit Ondřej Lhoták Objectives Spak is a modula toolkit fo flow-insensitive may points-to analyses fo Java, which enables expeimentation with: vaious paametes of pointe analyses which

More information

Lecture 27: Voronoi Diagrams

Lecture 27: Voronoi Diagrams We say that two points u, v Y ae in the same connected component of Y if thee is a path in R N fom u to v such that all the points along the path ae in the set Y. (Thee ae two connected components in the

More information

Module 6 STILL IMAGE COMPRESSION STANDARDS

Module 6 STILL IMAGE COMPRESSION STANDARDS Module 6 STILL IMAE COMPRESSION STANDARDS Lesson 17 JPE-2000 Achitectue and Featues Instuctional Objectives At the end of this lesson, the students should be able to: 1. State the shotcomings of JPE standad.

More information

Segmentation of Casting Defects in X-Ray Images Based on Fractal Dimension

Segmentation of Casting Defects in X-Ray Images Based on Fractal Dimension 17th Wold Confeence on Nondestuctive Testing, 25-28 Oct 2008, Shanghai, China Segmentation of Casting Defects in X-Ray Images Based on Factal Dimension Jue WANG 1, Xiaoqin HOU 2, Yufang CAI 3 ICT Reseach

More information

On the Conversion between Binary Code and Binary-Reflected Gray Code on Boolean Cubes

On the Conversion between Binary Code and Binary-Reflected Gray Code on Boolean Cubes On the Convesion between Binay Code and BinayReflected Gay Code on Boolean Cubes The Havad community has made this aticle openly available. Please shae how this access benefits you. You stoy mattes Citation

More information

DYNAMIC STORAGE ALLOCATION. Hanan Samet

DYNAMIC STORAGE ALLOCATION. Hanan Samet ds0 DYNAMIC STORAGE ALLOCATION Hanan Samet Compute Science Depatment and Cente fo Automation Reseach and Institute fo Advanced Compute Studies Univesity of Mayland College Pak, Mayland 074 e-mail: hjs@umiacs.umd.edu

More information

Detection and Recognition of Alert Traffic Signs

Detection and Recognition of Alert Traffic Signs Detection and Recognition of Alet Taffic Signs Chia-Hsiung Chen, Macus Chen, and Tianshi Gao 1 Stanfod Univesity Stanfod, CA 9305 {echchen, macuscc, tianshig}@stanfod.edu Abstact Taffic signs povide dives

More information

Scaling Location-based Services with Dynamically Composed Location Index

Scaling Location-based Services with Dynamically Composed Location Index Scaling Location-based Sevices with Dynamically Composed Location Index Bhuvan Bamba, Sangeetha Seshadi and Ling Liu Distibuted Data Intensive Systems Laboatoy (DiSL) College of Computing, Geogia Institute

More information

Point-Biserial Correlation Analysis of Fuzzy Attributes

Point-Biserial Correlation Analysis of Fuzzy Attributes Appl Math Inf Sci 6 No S pp 439S-444S (0 Applied Mathematics & Infomation Sciences An Intenational Jounal @ 0 NSP Natual Sciences Publishing o Point-iseial oelation Analysis of Fuzzy Attibutes Hao-En hueh

More information

GARBAGE COLLECTION METHODS. Hanan Samet

GARBAGE COLLECTION METHODS. Hanan Samet gc0 GARBAGE COLLECTION METHODS Hanan Samet Compute Science Depatment and Cente fo Automation Reseach and Institute fo Advanced Compute Studies Univesity of Mayland College Pak, Mayland 07 e-mail: hjs@umiacs.umd.edu

More information

An Extension to the Local Binary Patterns for Image Retrieval

An Extension to the Local Binary Patterns for Image Retrieval , pp.81-85 http://x.oi.og/10.14257/astl.2014.45.16 An Extension to the Local Binay Pattens fo Image Retieval Zhize Wu, Yu Xia, Shouhong Wan School of Compute Science an Technology, Univesity of Science

More information

A Two-stage and Parameter-free Binarization Method for Degraded Document Images

A Two-stage and Parameter-free Binarization Method for Degraded Document Images A Two-stage and Paamete-fee Binaization Method fo Degaded Document Images Yung-Hsiang Chiu 1, Kuo-Liang Chung 1, Yong-Huai Huang 2, Wei-Ning Yang 3, Chi-Huang Liao 4 1 Depatment of Compute Science and

More information

A New and Efficient 2D Collision Detection Method Based on Contact Theory Xiaolong CHENG, Jun XIAO a, Ying WANG, Qinghai MIAO, Jian XUE

A New and Efficient 2D Collision Detection Method Based on Contact Theory Xiaolong CHENG, Jun XIAO a, Ying WANG, Qinghai MIAO, Jian XUE 5th Intenational Confeence on Advanced Mateials and Compute Science (ICAMCS 2016) A New and Efficient 2D Collision Detection Method Based on Contact Theoy Xiaolong CHENG, Jun XIAO a, Ying WANG, Qinghai

More information

Towards Adaptive Information Merging Using Selected XML Fragments

Towards Adaptive Information Merging Using Selected XML Fragments Towads Adaptive Infomation Meging Using Selected XML Fagments Ho-Lam Lau and Wilfed Ng Depatment of Compute Science and Engineeing, The Hong Kong Univesity of Science and Technology, Hong Kong {lauhl,

More information

Optical Flow for Large Motion Using Gradient Technique

Optical Flow for Large Motion Using Gradient Technique SERBIAN JOURNAL OF ELECTRICAL ENGINEERING Vol. 3, No. 1, June 2006, 103-113 Optical Flow fo Lage Motion Using Gadient Technique Md. Moshaof Hossain Sake 1, Kamal Bechkoum 2, K.K. Islam 1 Abstact: In this

More information

Information Retrieval. CS630 Representing and Accessing Digital Information. IR Basics. User Task. Basic IR Processes

Information Retrieval. CS630 Representing and Accessing Digital Information. IR Basics. User Task. Basic IR Processes CS630 Repesenting and Accessing Digital Infomation Infomation Retieval: Basics Thosten Joachims Conell Univesity Infomation Retieval Basics Retieval Models Indexing and Pepocessing Data Stuctues ~ 4 lectues

More information

ANALYTIC PERFORMANCE MODELS FOR SINGLE CLASS AND MULTIPLE CLASS MULTITHREADED SOFTWARE SERVERS

ANALYTIC PERFORMANCE MODELS FOR SINGLE CLASS AND MULTIPLE CLASS MULTITHREADED SOFTWARE SERVERS ANALYTIC PERFORMANCE MODELS FOR SINGLE CLASS AND MULTIPLE CLASS MULTITHREADED SOFTWARE SERVERS Daniel A Menascé Mohamed N Bennani Dept of Compute Science Oacle, Inc Geoge Mason Univesity 1211 SW Fifth

More information

SYSTEM LEVEL REUSE METRICS FOR OBJECT ORIENTED SOFTWARE : AN ALTERNATIVE APPROACH

SYSTEM LEVEL REUSE METRICS FOR OBJECT ORIENTED SOFTWARE : AN ALTERNATIVE APPROACH I J C A 7(), 202 pp. 49-53 SYSTEM LEVEL REUSE METRICS FOR OBJECT ORIENTED SOFTWARE : AN ALTERNATIVE APPROACH Sushil Goel and 2 Rajesh Vema Associate Pofesso, Depatment of Compute Science, Dyal Singh College,

More information

Automatically Testing Interacting Software Components

Automatically Testing Interacting Software Components Automatically Testing Inteacting Softwae Components Leonad Gallaghe Infomation Technology Laboatoy National Institute of Standads and Technology Gaithesbug, MD 20899, USA lgallaghe@nist.gov Jeff Offutt

More information

ADDING REALISM TO SOURCE CHARACTERIZATION USING A GENETIC ALGORITHM

ADDING REALISM TO SOURCE CHARACTERIZATION USING A GENETIC ALGORITHM ADDING REALISM TO SOURCE CHARACTERIZATION USING A GENETIC ALGORITHM Luna M. Rodiguez*, Sue Ellen Haupt, and Geoge S. Young Depatment of Meteoology and Applied Reseach Laboatoy The Pennsylvania State Univesity,

More information

Assessment of Track Sequence Optimization based on Recorded Field Operations

Assessment of Track Sequence Optimization based on Recorded Field Operations Assessment of Tack Sequence Optimization based on Recoded Field Opeations Matin A. F. Jensen 1,2,*, Claus G. Søensen 1, Dionysis Bochtis 1 1 Aahus Univesity, Faculty of Science and Technology, Depatment

More information

Communication vs Distributed Computation: an alternative trade-off curve

Communication vs Distributed Computation: an alternative trade-off curve Communication vs Distibuted Computation: an altenative tade-off cuve Yahya H. Ezzeldin, Mohammed amoose, Chistina Fagouli Univesity of Califonia, Los Angeles, CA 90095, USA, Email: {yahya.ezzeldin, mkamoose,

More information

Also available at ISSN (printed edn.), ISSN (electronic edn.) ARS MATHEMATICA CONTEMPORANEA 3 (2010)

Also available at  ISSN (printed edn.), ISSN (electronic edn.) ARS MATHEMATICA CONTEMPORANEA 3 (2010) Also available at http://amc.imfm.si ISSN 1855-3966 (pinted edn.), ISSN 1855-3974 (electonic edn.) ARS MATHEMATICA CONTEMPORANEA 3 (2010) 109 120 Fulleene patches I Jack E. Gave Syacuse Univesity, Depatment

More information

Illumination methods for optical wear detection

Illumination methods for optical wear detection Illumination methods fo optical wea detection 1 J. Zhang, 2 P.P.L.Regtien 1 VIMEC Applied Vision Technology, Coy 43, 5653 LC Eindhoven, The Nethelands Email: jianbo.zhang@gmail.com 2 Faculty Electical

More information

arxiv: v4 [cs.ds] 7 Feb 2018

arxiv: v4 [cs.ds] 7 Feb 2018 Dynamic DFS in Undiected Gaphs: beaking the O(m) baie Suende Baswana Sheejit Ray Chaudhuy Keeti Choudhay Shahbaz Khan axiv:1502.02481v4 [cs.ds] 7 Feb 2018 Depth fist seach (DFS) tee is a fundamental data

More information

Cardiac C-Arm CT. SNR Enhancement by Combining Multiple Retrospectively Motion Corrected FDK-Like Reconstructions

Cardiac C-Arm CT. SNR Enhancement by Combining Multiple Retrospectively Motion Corrected FDK-Like Reconstructions Cadiac C-Am CT SNR Enhancement by Combining Multiple Retospectively Motion Coected FDK-Like Reconstuctions M. Pümme 1, L. Wigstöm 2,3, R. Fahig 2, G. Lauitsch 4, J. Honegge 1 1 Institute of Patten Recognition,

More information

An Optimised Density Based Clustering Algorithm

An Optimised Density Based Clustering Algorithm Intenational Jounal of Compute Applications (0975 8887) Volume 6 No.9, Septembe 010 An Optimised Density Based Clusteing Algoithm J. Hencil Pete Depatment of Compute Science St. Xavie s College, Palayamkottai,

More information

FACE VECTORS OF FLAG COMPLEXES

FACE VECTORS OF FLAG COMPLEXES FACE VECTORS OF FLAG COMPLEXES ANDY FROHMADER Abstact. A conjectue of Kalai and Eckhoff that the face vecto of an abitay flag complex is also the face vecto of some paticula balanced complex is veified.

More information

= dv 3V (r + a 1) 3 r 3 f(r) = 1. = ( (r + r 2

= dv 3V (r + a 1) 3 r 3 f(r) = 1. = ( (r + r 2 Random Waypoint Model in n-dimensional Space Esa Hyytiä and Joma Vitamo Netwoking Laboatoy, Helsinki Univesity of Technology, Finland Abstact The andom waypoint model (RWP) is one of the most widely used

More information

n If S is in convex position, then thee ae exactly k convex k-gons detemined by subsets of S. In geneal, howeve, S may detemine fa fewe convex k-gons.

n If S is in convex position, then thee ae exactly k convex k-gons detemined by subsets of S. In geneal, howeve, S may detemine fa fewe convex k-gons. Counting Convex Polygons in Plana Point Sets Joseph S. B. Mitchell a;1, Günte Rote b, Gopalakishnan Sundaam c, and Gehad Woeginge b a Applied Mathematics and Statistics, SUNY Stony Book, NY 11794-3600.

More information

A Novel Automatic White Balance Method For Digital Still Cameras

A Novel Automatic White Balance Method For Digital Still Cameras A Novel Automatic White Balance Method Fo Digital Still Cameas Ching-Chih Weng 1, Home Chen 1,2, and Chiou-Shann Fuh 3 Depatment of Electical Engineeing, 2 3 Gaduate Institute of Communication Engineeing

More information

DEADLOCK AVOIDANCE IN BATCH PROCESSES. M. Tittus K. Åkesson

DEADLOCK AVOIDANCE IN BATCH PROCESSES. M. Tittus K. Åkesson DEADLOCK AVOIDANCE IN BATCH PROCESSES M. Tittus K. Åkesson Univesity College Boås, Sweden, e-mail: Michael.Tittus@hb.se Chalmes Univesity of Technology, Gothenbug, Sweden, e-mail: ka@s2.chalmes.se Abstact:

More information

INDEXATION OF WEB PAGES BASED ON THEIR VISUAL RENDERING

INDEXATION OF WEB PAGES BASED ON THEIR VISUAL RENDERING INDEXATION OF WEB PAGES BASED ON THEIR VISUAL RENDERING Emmanuel Buno Univesité du Sud Toulon-Va / LSIS CNRS BP 20132, F-83957 La Gade buno@univ-tln.f Nicolas Faessel LSIS CNRS Domaine Univesitaie de Saint-Jéôme

More information

Any modern computer system will incorporate (at least) two levels of storage:

Any modern computer system will incorporate (at least) two levels of storage: 1 Any moden compute system will incopoate (at least) two levels of stoage: pimay stoage: andom access memoy (RAM) typical capacity 32MB to 1GB cost pe MB $3. typical access time 5ns to 6ns bust tansfe

More information

A VECTOR PERTURBATION APPROACH TO THE GENERALIZED AIRCRAFT SPARE PARTS GROUPING PROBLEM

A VECTOR PERTURBATION APPROACH TO THE GENERALIZED AIRCRAFT SPARE PARTS GROUPING PROBLEM Accepted fo publication Intenational Jounal of Flexible Automation and Integated Manufactuing. A VECTOR PERTURBATION APPROACH TO THE GENERALIZED AIRCRAFT SPARE PARTS GROUPING PROBLEM Nagiza F. Samatova,

More information

TESSELLATIONS. This is a sample (draft) chapter from: MATHEMATICAL OUTPOURINGS. Newsletters and Musings from the St. Mark s Institute of Mathematics

TESSELLATIONS. This is a sample (draft) chapter from: MATHEMATICAL OUTPOURINGS. Newsletters and Musings from the St. Mark s Institute of Mathematics TESSELLATIONS This is a sample (daft) chapte fom: MATHEMATICAL OUTPOURINGS Newslettes and Musings fom the St. Mak s Institute of Mathematics James Tanton www.jamestanton.com This mateial was and can still

More information

Spiral Recognition Methodology and Its Application for Recognition of Chinese Bank Checks

Spiral Recognition Methodology and Its Application for Recognition of Chinese Bank Checks Spial Recognition Methodology and Its Application fo Recognition of Chinese Bank Checks Hanshen Tang 1, Emmanuel Augustin 2, Ching Y. Suen 1, Olivie Baet 2, Mohamed Cheiet 3 1 Cente fo Patten Recognition

More information

Positioning of a robot based on binocular vision for hand / foot fusion Long Han

Positioning of a robot based on binocular vision for hand / foot fusion Long Han 2nd Intenational Confeence on Advances in Mechanical Engineeing and Industial Infomatics (AMEII 26) Positioning of a obot based on binocula vision fo hand / foot fusion Long Han Compute Science and Technology,

More information

Conservation Law of Centrifugal Force and Mechanism of Energy Transfer Caused in Turbomachinery

Conservation Law of Centrifugal Force and Mechanism of Energy Transfer Caused in Turbomachinery Poceedings of the 4th WSEAS Intenational Confeence on luid Mechanics and Aeodynamics, Elounda, Geece, August 1-3, 006 (pp337-34) Consevation Law of Centifugal oce and Mechanism of Enegy Tansfe Caused in

More information

Extract Object Boundaries in Noisy Images using Level Set. Final Report

Extract Object Boundaries in Noisy Images using Level Set. Final Report Extact Object Boundaies in Noisy Images using Level Set by: Quming Zhou Final Repot Submitted to Pofesso Bian Evans EE381K Multidimensional Digital Signal Pocessing May 10, 003 Abstact Finding object contous

More information

zeo, i.e., each watchman emains stationay, and we have the at galley poblem. At the othe limit, if m = 1, we have the watchman oute poblem. Anothe con

zeo, i.e., each watchman emains stationay, and we have the at galley poblem. At the othe limit, if m = 1, we have the watchman oute poblem. Anothe con Optimum Guad Coves and m-watchmen Routes fo Resticted Polygons Svante Calsson y Bengt J. Nilsson yz Simeon Ntafos x Abstact A watchman, in the teminology of at galleies, is a mobile guad. We conside seveal

More information

Obstacle Avoidance of Autonomous Mobile Robot using Stereo Vision Sensor

Obstacle Avoidance of Autonomous Mobile Robot using Stereo Vision Sensor Obstacle Avoidance of Autonomous Mobile Robot using Steeo Vision Senso Masako Kumano Akihisa Ohya Shin ichi Yuta Intelligent Robot Laboatoy Univesity of Tsukuba, Ibaaki, 35-8573 Japan E-mail: {masako,

More information

IP Multicast Simulation in OPNET

IP Multicast Simulation in OPNET IP Multicast Simulation in OPNET Xin Wang, Chien-Ming Yu, Henning Schulzinne Paul A. Stipe Columbia Univesity Reutes Depatment of Compute Science 88 Pakway Dive South New Yok, New Yok Hauppuage, New Yok

More information

Experimental and numerical simulation of the flow over a spillway

Experimental and numerical simulation of the flow over a spillway Euopean Wate 57: 253-260, 2017. 2017 E.W. Publications Expeimental and numeical simulation of the flow ove a spillway A. Seafeim *, L. Avgeis, V. Hissanthou and K. Bellos Depatment of Civil Engineeing,

More information

IP Network Design by Modified Branch Exchange Method

IP Network Design by Modified Branch Exchange Method Received: June 7, 207 98 IP Netwok Design by Modified Banch Method Kaiat Jaoenat Natchamol Sichumoenattana 2* Faculty of Engineeing at Kamphaeng Saen, Kasetsat Univesity, Thailand 2 Faculty of Management

More information

Prof. Feng Liu. Fall /17/2016

Prof. Feng Liu. Fall /17/2016 Pof. Feng Liu Fall 26 http://www.cs.pdx.edu/~fliu/couses/cs447/ /7/26 Last time Compositing NPR 3D Gaphics Toolkits Tansfomations 2 Today 3D Tansfomations The Viewing Pipeline Mid-tem: in class, Nov. 2

More information

(a, b) x y r. For this problem, is a point in the - coordinate plane and is a positive number.

(a, b) x y r. For this problem, is a point in the - coordinate plane and is a positive number. Illustative G-C Simila cicles Alignments to Content Standads: G-C.A. Task (a, b) x y Fo this poblem, is a point in the - coodinate plane and is a positive numbe. a. Using a tanslation and a dilation, show

More information

POMDP: Introduction to Partially Observable Markov Decision Processes Hossein Kamalzadeh, Michael Hahsler

POMDP: Introduction to Partially Observable Markov Decision Processes Hossein Kamalzadeh, Michael Hahsler POMDP: Intoduction to Patially Obsevable Makov Decision Pocesses Hossein Kamalzadeh, Michael Hahsle 2019-01-02 The R package pomdp povides an inteface to pomdp-solve, a solve (witten in C) fo Patially

More information

Combinatorial Mobile IP: A New Efficient Mobility Management Using Minimized Paging and Local Registration in Mobile IP Environments

Combinatorial Mobile IP: A New Efficient Mobility Management Using Minimized Paging and Local Registration in Mobile IP Environments Wieless Netwoks 0, 3 32, 200 200 Kluwe Academic Publishes. Manufactued in The Nethelands. Combinatoial Mobile IP: A New Efficient Mobility Management Using Minimized Paging and Local Registation in Mobile

More information

Prioritized Traffic Recovery over GMPLS Networks

Prioritized Traffic Recovery over GMPLS Networks Pioitized Taffic Recovey ove GMPLS Netwoks 2005 IEEE. Pesonal use of this mateial is pemitted. Pemission fom IEEE mu be obtained fo all othe uses in any cuent o futue media including epinting/epublishing

More information

Conversion Functions for Symmetric Key Ciphers

Conversion Functions for Symmetric Key Ciphers Jounal of Infomation Assuance and Secuity 2 (2006) 41 50 Convesion Functions fo Symmetic Key Ciphes Deba L. Cook and Angelos D. Keomytis Depatment of Compute Science Columbia Univesity, mail code 0401

More information

View Synthesis using Depth Map for 3D Video

View Synthesis using Depth Map for 3D Video View Synthesis using Depth Map fo 3D Video Cheon Lee and Yo-Sung Ho Gwangju Institute of Science and Technology (GIST) 1 Oyong-dong, Buk-gu, Gwangju, 500-712, Republic of Koea E-mail: {leecheon, hoyo}@gist.ac.k

More information

A modal estimation based multitype sensor placement method

A modal estimation based multitype sensor placement method A modal estimation based multitype senso placement method *Xue-Yang Pei 1), Ting-Hua Yi 2) and Hong-Nan Li 3) 1),)2),3) School of Civil Engineeing, Dalian Univesity of Technology, Dalian 116023, China;

More information

Slotted Random Access Protocol with Dynamic Transmission Probability Control in CDMA System

Slotted Random Access Protocol with Dynamic Transmission Probability Control in CDMA System Slotted Random Access Potocol with Dynamic Tansmission Pobability Contol in CDMA System Intaek Lim 1 1 Depatment of Embedded Softwae, Busan Univesity of Foeign Studies, itlim@bufs.ac.k Abstact In packet

More information

And Ph.D. Candidate of Computer Science, University of Putra Malaysia 2 Faculty of Computer Science and Information Technology,

And Ph.D. Candidate of Computer Science, University of Putra Malaysia 2 Faculty of Computer Science and Information Technology, (IJCSIS) Intenational Jounal of Compute Science and Infomation Secuity, Efficient Candidacy Reduction Fo Fequent Patten Mining M.H Nadimi-Shahaki 1, Nowati Mustapha 2, Md Nasi B Sulaiman 2, Ali B Mamat

More information

HISTOGRAMS are an important statistic reflecting the

HISTOGRAMS are an important statistic reflecting the JOURNAL OF L A T E X CLASS FILES, VOL. 14, NO. 8, AUGUST 2015 1 D 2 HistoSketch: Disciminative and Dynamic Similaity-Peseving Sketching of Steaming Histogams Dingqi Yang, Bin Li, Laua Rettig, and Philippe

More information

Embeddings into Crossed Cubes

Embeddings into Crossed Cubes Embeddings into Cossed Cubes Emad Abuelub *, Membe, IAENG Abstact- The hypecube paallel achitectue is one of the most popula inteconnection netwoks due to many of its attactive popeties and its suitability

More information

Layered Animation using Displacement Maps

Layered Animation using Displacement Maps Layeed Animation using Displacement Maps Raymond Smith, Wei Sun, Adian Hilton and John Illingwoth Cente fo Vision, Speech and Signal Pocessing Univesity of Suey, Guildfod GU25XH, UK a.hilton@suey.ac.uk

More information

A Mathematical Implementation of a Global Human Walking Model with Real-Time Kinematic Personification by Boulic, Thalmann and Thalmann.

A Mathematical Implementation of a Global Human Walking Model with Real-Time Kinematic Personification by Boulic, Thalmann and Thalmann. A Mathematical Implementation of a Global Human Walking Model with Real-Time Kinematic Pesonification by Boulic, Thalmann and Thalmann. Mashall Badley National Cente fo Physical Acoustics Univesity of

More information

A Minutiae-based Fingerprint Matching Algorithm Using Phase Correlation

A Minutiae-based Fingerprint Matching Algorithm Using Phase Correlation A Minutiae-based Fingepint Matching Algoithm Using Phase Coelation Autho Chen, Weiping, Gao, Yongsheng Published 2007 Confeence Title Digital Image Computing: Techniques and Applications DOI https://doi.og/10.1109/dicta.2007.4426801

More information

17/5/2009. Introduction

17/5/2009. Introduction 7/5/9 Steeo Imaging Intoduction Eample of Human Vision Peception of Depth fom Left and ight eye images Diffeence in elative position of object in left and ight eyes. Depth infomation in the views?? 7/5/9

More information

Query Language #1/3: Relational Algebra Pure, Procedural, and Set-oriented

Query Language #1/3: Relational Algebra Pure, Procedural, and Set-oriented Quey Language #1/3: Relational Algeba Pue, Pocedual, and Set-oiented To expess a quey, we use a set of opeations. Each opeation takes one o moe elations as input paamete (set-oiented). Since each opeation

More information

A General Characterization of Representing and Determining Fuzzy Spatial Relations

A General Characterization of Representing and Determining Fuzzy Spatial Relations 7 The Intenational Aab Jounal of Infomation Technolog A Geneal Chaacteization of Repesenting and Detemining Fuzz Spatial Relations Lui Bai and Li Yan 2 College of Infomation Science and Engineeing, Notheasten

More information

Desired Attitude Angles Design Based on Optimization for Side Window Detection of Kinetic Interceptor *

Desired Attitude Angles Design Based on Optimization for Side Window Detection of Kinetic Interceptor * Poceedings of the 7 th Chinese Contol Confeence July 6-8, 008, Kunming,Yunnan, China Desied Attitude Angles Design Based on Optimization fo Side Window Detection of Kinetic Intecepto * Zhu Bo, Quan Quan,

More information

Modelling, simulation, and performance analysis of a CAN FD system with SAE benchmark based message set

Modelling, simulation, and performance analysis of a CAN FD system with SAE benchmark based message set Modelling, simulation, and pefomance analysis of a CAN FD system with SAE benchmak based message set Mahmut Tenuh, Panagiotis Oikonomidis, Peiklis Chachalakis, Elias Stipidis Mugla S. K. Univesity, TR;

More information

Finding point-pairs. Find Closest Point from Dense Cloud

Finding point-pairs. Find Closest Point from Dense Cloud Finding point-pais Given an a, find a coesponding b on the suface. Then one appoach would be to seach evey possible tiangle o suface point and then take the closest point. The key is to find a moe efficient

More information

Parametric Query Optimization for Linear and Piecewise Linear Cost Functions

Parametric Query Optimization for Linear and Piecewise Linear Cost Functions Paametic Quey Oimization fo Linea and Piecewise Linea Cost Functions Avind Hulgei S. Sudashan Indian Institute of Technology, Bombay {au, sudasha}@cse.iitb.ac.in Abstact The of a quey an depends on many

More information

MapReduce Optimizations and Algorithms 2015 Professor Sasu Tarkoma

MapReduce Optimizations and Algorithms 2015 Professor Sasu Tarkoma apreduce Optimizations and Algoithms 2015 Pofesso Sasu Takoma www.cs.helsinki.fi Optimizations Reduce tasks cannot stat befoe the whole map phase is complete Thus single slow machine can slow down the

More information

Fifth Wheel Modelling and Testing

Fifth Wheel Modelling and Testing Fifth heel Modelling and Testing en Masoy Mechanical Engineeing Depatment Floida Atlantic Univesity Boca aton, FL 4 Lois Malaptias IFMA Institut Fancais De Mechanique Advancee ampus De lemont Feand Les

More information

XFVHDL: A Tool for the Synthesis of Fuzzy Logic Controllers

XFVHDL: A Tool for the Synthesis of Fuzzy Logic Controllers XFVHDL: A Tool fo the Synthesis of Fuzzy Logic Contolles E. Lago, C. J. Jiménez, D. R. López, S. Sánchez-Solano and A. Baiga Instituto de Micoelectónica de Sevilla. Cento Nacional de Micoelectónica, Edificio

More information

A Novel Image-Based Rendering System With A Longitudinally Aligned Camera Array

A Novel Image-Based Rendering System With A Longitudinally Aligned Camera Array EUOGAPHICS 2 / A. de Sousa, J.C. Toes Shot Pesentations A Novel Image-Based endeing System With A Longitudinally Aligned Camea Aay Jiang Li, Kun Zhou, Yong Wang and Heung-Yeung Shum Micosoft eseach, China

More information

Binary Dissection: Variants & Applications

Binary Dissection: Variants & Applications Binay Dissection: Vaiants & Applications Shahid H. Bokhai Depatment of Electical Engineeing Univesity of Engineeing & Technology Lahoe, Pakistan Thomas W. Cockett ICASE NASA Langley Reseach Cente Hampton,

More information

A Shape-preserving Affine Takagi-Sugeno Model Based on a Piecewise Constant Nonuniform Fuzzification Transform

A Shape-preserving Affine Takagi-Sugeno Model Based on a Piecewise Constant Nonuniform Fuzzification Transform A Shape-peseving Affine Takagi-Sugeno Model Based on a Piecewise Constant Nonunifom Fuzzification Tansfom Felipe Fenández, Julio Gutiéez, Juan Calos Cespo and Gacián Tiviño Dep. Tecnología Fotónica, Facultad

More information

4.2. Co-terminal and Related Angles. Investigate

4.2. Co-terminal and Related Angles. Investigate .2 Co-teminal and Related Angles Tigonometic atios can be used to model quantities such as

More information

Clustering Interval-valued Data Using an Overlapped Interval Divergence

Clustering Interval-valued Data Using an Overlapped Interval Divergence Poc. of the 8th Austalasian Data Mining Confeence (AusDM'9) Clusteing Inteval-valued Data Using an Ovelapped Inteval Divegence Yongli Ren Yu-Hsn Liu Jia Rong Robet Dew School of Infomation Engineeing,

More information

Evaluation of Partial Path Queries on XML data

Evaluation of Partial Path Queries on XML data Evaluation of Patial Path Queies on XML data Stefanos Souldatos Dept of EE & CE, NTUA stef@dblab.ntua.g Theodoe Dalamagas Dept of EE & CE, NTUA dalamag@dblab.ntua.g Xiaoying Wu Dept. of CS, NJIT xw43@njit.edu

More information

Accurate Diffraction Efficiency Control for Multiplexed Volume Holographic Gratings. Xuliang Han, Gicherl Kim, and Ray T. Chen

Accurate Diffraction Efficiency Control for Multiplexed Volume Holographic Gratings. Xuliang Han, Gicherl Kim, and Ray T. Chen Accuate Diffaction Efficiency Contol fo Multiplexed Volume Hologaphic Gatings Xuliang Han, Gichel Kim, and Ray T. Chen Micoelectonic Reseach Cente Depatment of Electical and Compute Engineeing Univesity

More information

High performance CUDA based CNN image processor

High performance CUDA based CNN image processor High pefomance UDA based NN image pocesso GEORGE VALENTIN STOIA, RADU DOGARU, ELENA RISTINA STOIA Depatment of Applied Electonics and Infomation Engineeing Univesity Politehnica of Buchaest -3, Iuliu Maniu

More information

Reachable State Spaces of Distributed Deadlock Avoidance Protocols

Reachable State Spaces of Distributed Deadlock Avoidance Protocols Reachable State Spaces of Distibuted Deadlock Avoidance Potocols CÉSAR SÁNCHEZ and HENNY B. SIPMA Stanfod Univesity We pesent a family of efficient distibuted deadlock avoidance algoithms with applications

More information

A ROI Focusing Mechanism for Digital Cameras

A ROI Focusing Mechanism for Digital Cameras A ROI Focusing Mechanism fo Digital Cameas Chu-Hui Lee, Meng-Feng Lin, Chun-Ming Huang, and Chun-Wei Hsu Abstact With the development and application of digital technologies, the digital camea is moe popula

More information

The EigenRumor Algorithm for Ranking Blogs

The EigenRumor Algorithm for Ranking Blogs he EigenRumo Algoithm fo Ranking Blogs Ko Fujimua N Cybe Solutions Laboatoies N Copoation akafumi Inoue N Cybe Solutions Laboatoies N Copoation Masayuki Sugisaki N Resonant Inc. ABSRAC he advent of easy

More information

Efficient protection of many-to-one. communications

Efficient protection of many-to-one. communications Efficient potection of many-to-one communications Miklós Molná, Alexande Guitton, Benad Cousin, and Raymond Maie Iisa, Campus de Beaulieu, 35 042 Rennes Cedex, Fance Abstact. The dependability of a netwok

More information

Effects of Model Complexity on Generalization Performance of Convolutional Neural Networks

Effects of Model Complexity on Generalization Performance of Convolutional Neural Networks Effects of Model Complexity on Genealization Pefomance of Convolutional Neual Netwoks Tae-Jun Kim 1, Dongsu Zhang 2, and Joon Shik Kim 3 1 Seoul National Univesity, Seoul 151-742, Koea, E-mail: tjkim@bi.snu.ac.k

More information

Annales UMCS Informatica AI 2 (2004) UMCS

Annales UMCS Informatica AI 2 (2004) UMCS Pobane z czasopisma Annales AI- Infomatica http://ai.annales.umcs.pl Annales Infomatica AI 2 (2004) 33-340 Annales Infomatica Lublin-Polonia Sectio AI http://www.annales.umcs.lublin.pl/ Embedding as a

More information

Title. Author(s)NOMURA, K.; MOROOKA, S. Issue Date Doc URL. Type. Note. File Information

Title. Author(s)NOMURA, K.; MOROOKA, S. Issue Date Doc URL. Type. Note. File Information Title CALCULATION FORMULA FOR A MAXIMUM BENDING MOMENT AND THE TRIANGULAR SLAB WITH CONSIDERING EFFECT OF SUPPO UNIFORM LOAD Autho(s)NOMURA, K.; MOROOKA, S. Issue Date 2013-09-11 Doc URL http://hdl.handle.net/2115/54220

More information

Improvement of First-order Takagi-Sugeno Models Using Local Uniform B-splines 1

Improvement of First-order Takagi-Sugeno Models Using Local Uniform B-splines 1 Impovement of Fist-ode Takagi-Sugeno Models Using Local Unifom B-splines Felipe Fenández, Julio Gutiéez, Gacián Tiviño and Juan Calos Cespo Dep. Tecnología Fotónica, Facultad de Infomática Univesidad Politécnica

More information

Quality Aware Privacy Protection for Location-based Services

Quality Aware Privacy Protection for Location-based Services In Poceedings of the th Intenational Confeence on Database Systems fo Advanced Applications (DASFAA 007), Bangkok, Thailand, Apil 9-, 007. Quality Awae Pivacy Potection fo Location-based Sevices Zhen Xiao,,

More information

Extended Perspective Shadow Maps (XPSM) Vladislav Gusev, ,

Extended Perspective Shadow Maps (XPSM)   Vladislav Gusev, , Extended Pespective Shadow Maps (XPSM) http://xpsm.og Vladislav Gusev,.8.27, xmvlad@gmail.com Figue : XPSM esults (~4 objects in a scene, 536x536 shadow map). Intoduction Shadows ae one of the most impotant

More information

Evaluation of Partial Path Queries on XML Data

Evaluation of Partial Path Queries on XML Data Evaluation of Patial Path Queies on XML Data Stefanos Souldatos Dept of EE & CE NTUA, Geece stef@dblab.ntua.g Theodoe Dalamagas Dept of EE & CE NTUA, Geece dalamag@dblab.ntua.g Xiaoying Wu Dept. of CS

More information

Goal. Rendering Complex Scenes on Mobile Terminals or on the web. Rendering on Mobile Terminals. Rendering on Mobile Terminals. Walking through images

Goal. Rendering Complex Scenes on Mobile Terminals or on the web. Rendering on Mobile Terminals. Rendering on Mobile Terminals. Walking through images Goal Walking though s -------------------------------------------- Kadi Bouatouch IRISA Univesité de Rennes I, Fance Rendeing Comple Scenes on Mobile Teminals o on the web Rendeing on Mobile Teminals Rendeing

More information