A Novel Feature Line Segment Approach for Pattern Classification

Size: px
Start display at page:

Download "A Novel Feature Line Segment Approach for Pattern Classification"

Transcription

1 12th International Conference on Information Fuion Seattle, WA, USA, July 6-9, 2009 A Novel Feature Line Segment Approach for Pattern Claification Yi Yang Intitute of Integrated Automation Xi an Jiaotong Univerity Xi an, Shaanxi, China P.R. jiafeiyy@163.com Chongzhao Han Intitute of Integrated Automation Xi an Jiaotong Univerity Xi an, Shaanxi, China P.R. czhan@mail.xjtu.edu.cn Deqiang Han Intitute of Integrated Automation Xi an Jiaotong Univerity Xi an, Shaanxi, China P.R. deqhan@mail.xjtu.edu.cn Abtract - In thi paper, a novel pattern claification approach i propoed called hortet feature line egment (SFLS). It retain the idea and advantage of nearet feature line (NFL) and it can uppre the drawback of NFL, i.e., the extrapolation inaccuracy, interpolation inaccuracy and high computational complexity. SFLS ue length of the feature line egment atifying given geometric relation contraint, intead of the perpendicular ditance from query point to feature line in NFL. SFLS ha clear geometric-theory foundation and it implementation i relatively imple. In experiment baed on artificial dataet and real-world dataet, comparion between SFLS and other claification method are provided, including nearet neighbor (NN), k-nn, NFL and ome refined NFL method. Experimental reult how that SFLS i a imple yet effective claification approach. Keyword: Claification, nearet neighbor (NN), nearet feature line (NFL), interpolation inaccuracy and extrapolation inaccuracy. 1 Introduction With the development of pattern claification[1], variou claification method have been propoed till now. The nearet neighbor (NN) rule i a non-parametric claification method which i imple, yet effective. It ha been proved that NN ha aymptotic error rate that i at mot twice the Baye error rate [2]. NN alo ha it own drawback, i.e., the repreentational capacity of the dataet and the error rate depend on how the training ample are choen to account for poible variation and alo how many training ample are available [3]. Lot of approache have been propoed which aim to counteract the drawback. k-nn, urrounding neighbor (SN) [4], graphic neighbor (GN) [5] and nearet feature line (NFL) [3,6] are all repreentative. NFL i a non-parametrilaifier which attempt to improve the repreentational capacity of a ample et with limited ize by uing the traight line which pa through each pair of the ample from the ame cla. A good argument for doing thi i that it add extra information to the ample et. NFL ha hown wonderful performance in many application, including face recognition [3], audio retrieval [7], peaker identification [8], image claification [6], etc. Though ucceful in improving the claification ability, there are alo drawback in NFL. Two kind of trepaing can caue two type of inaccuracie: extrapolation inaccuracy and interpolation inaccuracy [9,10]. In addition, NFL ha relatively high computational complexity. A mentioned by the author of [9]: extrainformation i a double-edged word. There are everal method propoed to rectify the NFL to counteract it drawback. In [11], center-baed nearet neighbor (CNN) method wa propoed to leen the computation burden of the original NFL method by defining another kind of line called center-baed line (CL). CL connect a ample point with known label and the center of the ample correponding cla, intead of two labeled ample point in the NFL method. In [10], the author pointed out one of the inaccuracie referred above extrapolation inaccuracy, and propoed a olution called nearet neighbor line (NNL). NNL alo can ignificantly reduce the computation time. In[12], a tunable nearet neighbor (TNN) method wa propoed to improve the performance of NFL. In [13], nearet feature midpoint (NFM) wa propoed to refine NFL by defining the ditance metric a the minimum Euclidean ditance between the query point and the feature midpoint. The computation complexity of NFM i ignificantly le than NFL. All thee refined method are effective and rational, but they till have their own drawback, i.e. they can not thoroughly reolve the problem of extrapolation inaccuracy and interpolation inaccuracy. Rectified nearet feature line egment (RNFLS) wa propoed in [9] to overcome both of the two inaccuracie mentioned above. Thu it can ignificantly improve the claification performance when compared to NFL. RNFLS alo ha it deficiencie which are dicued in next ection. In thi paper, we propoe a imple yet effective approach called hortet feature line egment (SFLS), which can uppre both the extrapolation inaccuracy and interpolation inaccuracy. Our approach ha clear geometrictheory foundation and retain the idea and advantage of original NFL, i.e., it ue a linear model of pair of ample point within the ame cla. It ue line feature intead of ISIF 490

2 point feature. Intead of calculating the ditance between the query point and the feature line, SFLS attempt to find the hortet feature line egment which atifie given geometric relation requirement. Experimental reult how the wonderful performance of SFLS in artificial dataet and real-world dataet. Comparion between SFLS and ome other claification method uch a NN, k-nn, NFL and refined NFL are alo provided to verify the rationality and efficacy of SFLS propoed. 2 Nearet feature line method The nearet feature line (NFL) method i deigned to utilize the information provided by each pair of point in the ame cla by contituting ome feature line (FL) pace. The NFL ditance i defined a the Euclidean ditance from the query point to the FL, i.e., the ditance between a query point and it projection onto the FL a illutrated in Figure 1. NFL i conitently uperior to the NN method baed on conventional ditance [12]. Figure 1. Nearet feature line 2.1 Brief review of NFL Let x, x j be two training ample from the ame cla, = 1,..., M, where M repreent the number of cla. Each training ample feature vector ha n dimenion. xi x j denote the traight line paing through x i and x j, which i called a feature line (FL) of the cla. Let x q be the query point and x ij be the projection point of x q on FL x i x j. x ij can be calculated i a follow ( x q, x i and where x j are all column vector): xij = (1 µ ) xi + µ x j (1) T ( xq xi ) ( xj xi ) µ = T ( x x ) ( x x ) j i j i (2) The ditance from the query point x q to the FL xi x j can be calculated a: d ( x, x x ) = x x (3) q i j q ij where denote ome norm. The claification deciion can be done a follow: = arg min d (4) 1 M where i the cla label aigned to query point. dnfl = d i called the NFL ditance. The repreentational capacity of training ample can be generalized by NFL, i.e., NFL can add extra information to training ample. For a given training ample et, quantity of line feature baed on NFL i larger than that of point feature. However the extra information might be a doubleedged word. Though ucceful in improving the claification performance, there are till ome drawback in NFL that will limit it further application in practice. Two main drawback can be ummarized a follow: (a) It will encounter large computation complexity problem when there are too many ample in each cla [11]. If there are N training ample belong to cla, there will be N( N 1)/2feature line. Suppoe that each ample feature vector i n dimenion, for traditional NFL, there will be (3 n+ 1) ( N ( N 1)/2) multiplication operation. In [8], a fat algorithm to calculating d( xq, xi x j) i propoed, computation time of which i only 1/3of traditional one. (b) It will encounter the problem of extrapolation inaccuracy and interpolation inaccuracy, which will be dicued further afterward. 2.2 Extrapolation inaccuracy and interpolation inaccuracy of NFL In NFL, only perpendicular ditance between query ample and feature line i ued a the deciion criterion. When a traight line of one cla trepae into of other clae area, it may lead to claification error. Two type of trepaing will be dicued in thi ection, which can caue two type of inaccuracie: extrapolation inaccuracy and interpolation inaccuracy [9] Extrapolation inaccuracy In Figure 2, query ample x q i urrounded by ample belong to cla Star, but it i claified to cla Circle with the criterion of NFL. Thi claification error i due to the extrapolating part of FL xx

3 cla Circle. Thi i due to the interpolating part of FL xx 1 2. To overcome the interpolation inaccuracy, author in [9] propoed a method for ample election. They contruct rectified nearet feature line egment (RNFLS) ubpace to repreent each cla by having thoe line egment trepaing the cla-area of other clae removed. Figure 2. Extrapolation inaccuracy. Figure 4. Interpolation inaccuracy. Figure 3. Sample point with the ame NFL ditance. We can ee that, all the point at two parallel dah line have the ame NFL ditance according to FL x 1 x 2. All the query point with the ame NFL ditance will be dealt without dicrimination. Thi i an important reaon for generation of extrapolation inaccuracy. It can be proved that extrapolation inaccuracy can be ignored if the feature pace dimenion i large enough [9]. But in a feature pace of low dimenion, it actually harm. To counteract the extrapolation inaccuracy, the author in [10] propoed nearet neighbor line (NNL). In NNL, only one feature line linking the nearet two ample point to the query i ued to repreent the cla. NNL can avoid the extrapolation problem, but it may alo loe ome claification ability [9]. In [13], nearet feature midpoint (NFM) wa propoed which attempt to overcome the extrapolation inaccuracy by defining the ditance metric a the minimum Euclidean ditance between the query point and the feature line midpoint. In other reearche uch a tunable nearet neighbor (TNN) [12], rectified nearet feature line egment (RNFLS) [9] and extended nearet feature line(enfl) [14], they modified the calculation method of the NFL ditance metric baed on the relative poition relationhip between the query point and the feature line. Thu the query point with the ame original NFL ditance can be dealt with dicrimination and extrapolation inaccuracy can be counteracted Interpolation inaccuracy An example of interpolation inaccuracy in NFL i illutrated in Figure 4. The query ample x q i urrounded by the ample belong to cla Star, but xq i claified to Figure 5. Sample ditribution cauing ignificant interpolation and extrapolation inaccuracy problem. Figure 5 illutrate a cae which can produce both the extrapolation inaccuracy and interpolation inaccuracy. The refined NFL method including NNL, NFM, TNN, can only uppre the extrapolation inaccuracy. The RNFLS method can overcome both the extrapolation and interpolation inaccuracy, o it i an effective modified NFL claification method. It can implement claification well in the environment illutrated in Figure 5. In the ame environment, the claification error will increae ignificantly for NFL. However, RNFLS alo ha it drawback. It implementation i relatively complex. It ha the procedure of ample ubpace election and calculation method of the ditance metric are variou according to the relative poition relationhip between the query point and the feature line egment. It computational complexity i relatively high. In thi paper, we propoe a novel feature line egment approach, which i imple yet effective and can uppre both the interpolation inaccuracy and extrapolation inaccuracy. 3 Shortet feature line egment claification approach In thi ection, a novel claification approach called hortet feature line egment (SFLS) i introduced. SFLS doe not calculate the ditance between the query point and the feature line. Intead, it attempt to find the hortet feature line egment atifying given geometric relation 492

4 contraint. Two ample in the ame cla contitute a feature line egment. If the query point i inide or on the hyperphere centered at the midpoint of the feature line egment and whoe diameter i length of the feature line egment, the correponding feature line egment will be tagged. Then find out the hortet feature line egment in all the tagged feature line egment and aign the hortet feature line cla label to the query point. The pecific claification procedure i a follow: Let x i, x j be two training ample from the ame cla, = 1,..., M. M repreent number of cla. If there are N ample belonging to cla in training et, there will be N( N 1)/2 feature line egment. For a query point x q, execute the tep a follow: 1) Calculate the angle (denoted byα ) between vector: x q x i and x q x j baed on formula (5): T 180 ( xq xi ) ( xq xj) α = arcco (5) π x x x x Where q i q j x q, xi and x j are all column vector and denote the Euclidean norm. The unit of α i deg. If α i an acute angle (i.e., 0 α < 90 ), leave feature line egment xi x j untagged. If α i a right angle or an obtue angle (i.e., 90 α 180 ), tag feature line egment x i x j. Such feature line egment are called tagged feature line. 2) Find out the hortet tagged feature line egment: i j x x, label cla to query point x q. If there i no tagged feature line, then the correponding query point i rejected to be claified. Baed on the baic geometric theorem, a illutrated in Fig. 6, if α i an acute angle, then x q i outide the hyperphere; if α i a right angle, then x q i on the hyperphere; if α i an obtue angle, then x q i inide the hyperphere. The hyperphere center i the midpoint of the feature line egment xi x j and it diameter i the length of the feature line egment xi x j. (a) (b) (c) Figure 6. Relative poition relation between query point and feature line egment. (a) Inide hyperphere; (b) On hyperphere; Outide hyperphere. The claification procedure can be briefly by an example illutrated in Figure 7. Figure 7. Shortet feature line egment claification. Query point x q i inide hyperphere x q centered at midpoint of xx 1 2 and whoe diameter i x 1 x 2. x q i alo inide hyperphere C entered at midpoint of xx 1 2 and whoe diameter i x1 x 2. Feature line egment 1 2 xx and xx 1 2 are both tagged. xx 1 2 i horter than xx 1 2, o query point xq i labeled cla Circle. 4 Analyi of SFLS SFLS propoed in thi paper ha powerful claification ability and it can uppre ome ignificant drawback of NFL. SFLS i alo relatively eay to be implemented and it ha relatively low computational complexity. 4.1 Claification ability analyi (1) SFLS retain the idea and advantage of NFL SFLS ue a linear model of each pair of ample point within the ame cla to generalize the repreentational capacity of ample et. It ue line feature intead of point feature. (2)SFLS ha the concentration property of feature line egment SFLS ue the feature line egment intead of feature line. The feature pace contituted by feature line egment ha a good property, i.e., the concentration[9]. It can be concluded that tronger concentration property can bring the improvement of claification accuracy[9]. (3)SFLS ha more reaonable imilarity definition NFL only ue the perpendicular ditance from the query point to the feature line a the criterion for claification deciion. The perpendicular ditance i in fact the imilarity between point feature and line feature. What we concern i the imilarity between the query ample and the given cla. Only uch a ditance i not ufficient to repreent the imilarity between query ample and the feature line correponding cla. Le perpendicular ditance doe not alway repreent more imilarity. Such a propoition can be upported by the two kind of inaccuracie aforementioned. In SLFS, a referred above, 493

5 both the ditance and the volume of the hyperphere contituted by feature line egment and query point are emphaized to define the imilarity. For a tagged feature line egment, the horter the feature line egment i, the le the volume of the hyperphere i, which cover the query point and the feature line egment. Le volume repreent more imilarity between query point and feature line egment. According to the imilarity definition in SFLS, there will be a good property: the query point i alway near to both the two point contituting the hortet feature line egment. It can better reflect the imilarity between the query point and the correponding cla. (4)SFLS can uppre inaccuracie caued by trepa Firt, SFLS ue line egment intead of traight line. Intuitively, a line egment ha finite length while a traight line extend to infinity. Thu the probability to produce trepa problem can be reduced when uing feature line egment. Second, SFLS ha no problem of extrapolation inaccuracy. The poition parameter µ can be calculated. In SFLS, according to Formula (2), 0 µ 1definitely come into exitence. It mean that the projection of x q onto the feature line i an interpolating point of the tagged feature line egment. Third, in ome cae, the interpolation inaccuracy can be uppreed in SFLS a illutrated in Figure 8. Figure 8. NFL and SFLS in interpolation inaccuracy problem. For query point x q, the hortet feature line egment in cla circle i i j xx and in cla Star i xx. The query point x q urrounded by ample of cla Star i claified to cla Circle by NFL due to the c interpolation inaccuracy ( d < d ). But it i correctly claified to cla Star by SFLS becaue xx i j > xx i j. But it hould be noted that SFLS can not thoroughly reolve the interpolation inaccuracy. i j 4.2 Computational complexity analyi If there are N ample belong to cla in training et, we will have N( N 1)/2 feature line. Suppoe that feature vector of each ample ha n dimenion, the computation in SFLS for each cla include (3 n) ( N ( N 1)/2) multiplication operation which i le than that of the original NFL approach: (3 n+ 1) ( N ( N 1)/2). And it hould be noted that there i no offline ample ubpace election or preparation or the claification baed on SFLS. In RNFLS, offline preparation i needed, which ha computation time 3 complexity of ON ( ), where N denote the number of training ample. In general, SFLS propoed in thi paper i a imple yet effective claification method. It ha clear geometric theory foundation. It retain the main idea and advantage of NFL and at ame time it can counteract extrapolation inaccuracy and interpolation inaccuracy. SFLS attempt to find the hortet feature line egment which i near to the query point. However, the near i not defined baed on the perpendicular ditance from the query point to the feature line. It ue the geometric relation to define the near, which repreent the imilarity between the query and the feature line egment. 5 Experiment The claification performance of SFLS i compared with claification approache including NN, k-nn, traditional NFL, NNL, TNN, RNFLS. The experiment are executed on two artificial dataet and ome real-world dataet. 5.1 Artificial dataet: Two-piral The two piral curve can be generated baed on formula (6) and formula (7) a follow [9,15-17]: x = k co( ) piral1: (6) y = k in( ) x = k co( + π) piral2: (7) y = k in( + π) Set parameter π /2 3π and uppoe that each cla probability denity i uniform along the correponding curve. The two-piral dataet for experiment i polluted by Gauian noie whoe mean i zero and variance iσ = 1.5. The two piral curve and the polluted two piral curve for experiment are illutrated in Figure 9. Totally 500 ample are generated. Cla 1 and cla 2 each have 250 ample. Randomly elect 125 training ample from each cla, the remainder are reerved for tet. Then we ue NN, k-nn, NFL, NNL, RNFLS and the SFLS propoed to perform the claification. The k parameter in k-nn i et to 3. The claification procedure referred above i executed for 10 time. In each time, ample for training and tet are re- 494

6 elected randomly. The experimental reult are lited in Table 1. where µ denote the ditance between the two Gauian center. Totally 500 ample are generated. Cla 1 and cla 2 each have 250 ample. Randomly elect 125 training ample from each cla, the remainder are reerved a tet ample. 10 different µ value are ued: ( µ = 0.5, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 5.5 ). We ue NN, k-nn, NFL, NNL, RNFLS and the SFLS propoed to perform the claification. The k parameter in k-nn i et to 3. The experimental reult are hown in Figure 10 and Table 2. The Bayeian claifier ue the true ditribution (Guaian). Figure 9. Two-piral curve. (a) Two-piral; (b) Two-piral polluted with Gauian noie. Table 1 Two-Spiral Claification performance Claification Approach Correct Rate (Min value) Correct Rate (Max value) Correct Rate (Mean) NN 93.6% 96.8% 95.5% k-nn 94.8% 98.4% 96.6% NFL 52.0% 63.2% 56.0% NNL 85.2% 91.6% 89.2% RNFLS 94.0% 97.2% 96.0% SFLS 94.0% 97.2% 96.0% It can be een that NFL i not proper to be ued in two-piral claification problem. It i due to the interpolation inaccuracy and extrapolation inaccuracy referred above. RNFLS gain better performance due to the building of the RNFLS ubpace and the new definition of ditance metric baed on the relative poition relationhip between query point and feature line egment. SFLS propoed in thi paper i relatively imple but it alo achieve the ame claification accuracy a RNFLS. 5.2 Claification problem of two clae with Gauian ditribution To compare the claification performance of the SFLS and other claification method with that of Bayeian claifier, we generate the artificial two-cla dataet with known probability denity- Guaian Ditribution. Any ample in our dataet ha two dimenion ( x, y ). x i a Gauian ditributed variable and y i a uniform ditributed variable. Their joint-probability denity function i a follow [9]: 1 1 x 2 exp[ ( ) ], a y b, pcla1( x, y) = 2 πσ( b a) 2 σ 0, otherwie; 1 1 x µ 2 exp[ ( ) ], a y b, pcla2( x, y) = 2 πσ( b a) 2 σ 0, otherwie; (8) Figure 20. Claification performance of two-gauian ditribution ample. Table 2 Claification performance of two-gauian ditribution ample Claification approach Mean claification correct rate of experiment with variou ditance of Gauian center Bayeian 88.1% NN 83.1% 3-NN 85.0% NFL 69.7% NNL 80.8% RNFLS 85.7% SFLS 85.7% Baed on the reult, it can be concluded that the NFL i not fit for thi type of dataet. Thi i mainly due to the extrapolation inaccuracy. The NN, k-nn and the refined NFL approache can achieve relatively good claification performance. SFLS and RNFLS perform the bet except for Bayeian approach. However, SFLS i impler. 5.3 Claification problem baed on realworld dataet The artificial dataet ued above are two-cla. To verify the SFLS performance in multi-cla (cla number 3) claification problem, we ue three multiple-cla real- 495

7 world dataet from UCI [18] including Wine (3-cla), Iri (3-cla) and Gla (6-cla). For Wine and Gla dataet, all attribute value are normalized to lie between 0 and 1 baed on formula (9): vi min vi ai = (9) max vi min vi where v i denote the actual value of attribute i. Maximum and minimum operation are over all ample. Each dataet i randomly divided into 5 dijoint ubet of approximately equal ize. For each time, we elect three ubet to contitute the training et and treat the remainder a the teting et. Thu, there are totally C = different trial over each dataet. The average claification performance acro the 10 trial are computed. The k parameter in k-nn i et to 3. The experimental reult are illutrated in Figure 11. Baed on the experimental reult illutrated in Figure 11, we can conclude that the SFLS approach propoed in thi paper alway achieve better performance than the other over the 3 real-world dataet. Figure 11. Claification performance baed on three dataet from UCI. 5.4 Dicuion Claification baed on line feature are not definitely better than thoe baed on point feature. When the cale of ample et i limited and when the feature vector of ample point have relatively low dimenion, the advantage of line feature (or line egment feature) can be ditinctly hown, which can alo been upported by the provided experimental reult. There are everal NFL-refined claification approache ued in the experiment. SFLS and RNFLS alway perform better than original NFL and other NFL-refined method. When compared to other method baed on line feature, SFLS ha more rational and more comprehenive definition of imilarity, which i crucial for the claification problem. Compared to RNFLS, both SFLS and RNFLS ue feature line egment which ha the good property of concentration and they can uppre the inaccuracie caued by trepa, but SFLS doe not need the preparation procedure for ample ubpace. SFLS ue imply the length of feature line egment intead of variou ditance definition in RNFLS. SFLS i relatively imple yet effective which can do the claification at leat a well a RNFLS in average when the number of cla i relatively low and SFLS perform better than RNFLS when the number of cla i relatively high. Thee can be upported by the provided experimental reult. It hould be noted that for SFLS, in the wort cae where there i no tagged feature line for a tet ample, if rejection deciion i not permitted, jut ue the rule of nearet neighbor (NN) intead to make the claification deciion for the ample which are reject to be claified by SFLS. 6 Concluion In thi paper, a novel claification approach called SFLS i propoed. SFLS can be regarded a a refined NFL method which can ubtantially improve the claification performance of NFL. Theoretical jutification and analyi of SFLS rationality are provided. Experimental reult baed on artificial dataet and real-world dataet (including UCI, etc.) alo how that SFLS i an effective claification approach. In the future, we will do further reearch in SFLS. The computational cot of SFLS i till not a low a expected. Fater and more effective algorithm for SFLS are worth reearching. Some modification to SFLS can be made according to other effective claification approache. For example, baed on SFLS, k-sfls can be developed, jut like the idea of k-nn, better performance can be expected. In thi paper, we only dicu about the claification problem with Euclidean ditance. Other ditance definition can alo be ued to implement the SFLS approach. Thi can alo be our work in the future. Acknowledgement Thi work i upported in part by the National Natural Science Foundation of China, No and the Grant for State Key Program for Baic Reearch of China (973) No.2007CB and in part by National Natural Science Foundation of China (No ) and China Potdoctoral Science Foundation (No ). Reference [1] S. Theodoridi, and K. Koutroumba, Pattern Recognition, 2nd ed., Elevier, Singapore, [2] R. O. Duda, and P.E. Hart, D.G. Stork, Pattern Claification, 2nd ed., Wiley Inter-cience publication, New York, [3] S. Z. Li, and J. W. Lu, Face recognition uing the nearet feature line method, IEEE Tranaction on Neural Network, Vol. 10, No. 2, pp ,

8 [4] B. B. Chaudhuri, A new definition of neighborhood of a point in multi-dimenional pace, Pattern Recognition Letter, Vol. 17, No. 1, pp , [5] J. S. Sanchez, F. Pla, and F. J. Ferri, On the ue of neighbourhood-baed non-parametrilaifier, Pattern Recognition Letter, Vol. 18, No , pp , [6] S. Z. Li, K. L. Chan, and C. L. Wang, Performance evaluation of the nearet feature line method in image claification and retrieval, IEEE Tranaction on Pattern Analyi and Machine Intelligence, Vol.22, No. 11, pp , [7] S. Z. Li, Content-baed audio claification and retrieval uing the nearet feature line method," IEEE Tranaction on Speech and Audio Proceing, Vol. 8, No. 5, pp , [8] K. Chen, T. Y. Wu, and H. J. Zhang, On the ue of nearet feature line for peaker identification, Pattern Recognition Letter, Vol. 23, No. 14, pp , [9] H. Du, and Y. Q. Chen, Rectified nearet feature line egment for pattern claification, Pattern Recognition, Vol. 40, No. 5, pp , [10] W. Zheng, L. Zhao, and C. Zou, Locally nearet neighbor claifier for pattern claification, Pattern Recognition, Vol. 37, No. 6, pp , [11] Q. B. Gao, and Z. Z. Wang, Center-baed nearet neighbor claifier, Pattern Recognition, Vol. 40, No. 1, pp , [12] Y. Zhou, C. Zhang, and J. Wang, Tunable Nearet Neighbor Claifier, Pattern Recognition, Vol. 3175, Springer, Berlin, 2004, pp [13] Z. Zonglin, K. Chee Keong, The pattern claification baed on the nearet feature midpoint, in: K. Chee Keong (Ed.), Pattern Recognition, ICPR Proceeding of the 17th International Conference on, Vol. 3, pp [14] Y. Zhou, C. Zhang, and J. Wang, Extended Nearet Feature Line Claifier, PRICAI 2004: Trend in Artificial Intelligence, 2004, pp [15] T. Denoeux, and R. Lengelle, Initializing back propagation network with prototype, Neural Network, Vol. 6, No. 3, pp , [16] Y. Q. Chen, D.W. Thoma, and M.S. Nixon, Generating-hrinking algorithm for learning arbitrary claification, Neural Network, Vol. 7, No. 9, pp , [17] S. K. Sin, and R. J. P. DeFigueiredo, Efficient learning procedure for optimal interpolative net, Neural Network, Vol. 6, No. 1, pp , [18] C. L. BLAKE, and C. L. Merz, UCI repoitory of machine learning databae, (Online), available < 497

Lecture 14: Minimum Spanning Tree I

Lecture 14: Minimum Spanning Tree I COMPSCI 0: Deign and Analyi of Algorithm October 4, 07 Lecture 4: Minimum Spanning Tree I Lecturer: Rong Ge Scribe: Fred Zhang Overview Thi lecture we finih our dicuion of the hortet path problem and introduce

More information

Markov Random Fields in Image Segmentation

Markov Random Fields in Image Segmentation Preented at SSIP 2011, Szeged, Hungary Markov Random Field in Image Segmentation Zoltan Kato Image Proceing & Computer Graphic Dept. Univerity of Szeged Hungary Zoltan Kato: Markov Random Field in Image

More information

3D SMAP Algorithm. April 11, 2012

3D SMAP Algorithm. April 11, 2012 3D SMAP Algorithm April 11, 2012 Baed on the original SMAP paper [1]. Thi report extend the tructure of MSRF into 3D. The prior ditribution i modified to atify the MRF property. In addition, an iterative

More information

Routing Definition 4.1

Routing Definition 4.1 4 Routing So far, we have only looked at network without dealing with the iue of how to end information in them from one node to another The problem of ending information in a network i known a routing

More information

Trainable Context Model for Multiscale Segmentation

Trainable Context Model for Multiscale Segmentation Trainable Context Model for Multicale Segmentation Hui Cheng and Charle A. Bouman School of Electrical and Computer Engineering Purdue Univerity Wet Lafayette, IN 47907-1285 {hui, bouman}@ ecn.purdue.edu

More information

Minimum congestion spanning trees in bipartite and random graphs

Minimum congestion spanning trees in bipartite and random graphs Minimum congetion panning tree in bipartite and random graph M.I. Otrovkii Department of Mathematic and Computer Science St. John Univerity 8000 Utopia Parkway Queen, NY 11439, USA e-mail: otrovm@tjohn.edu

More information

Universität Augsburg. Institut für Informatik. Approximating Optimal Visual Sensor Placement. E. Hörster, R. Lienhart.

Universität Augsburg. Institut für Informatik. Approximating Optimal Visual Sensor Placement. E. Hörster, R. Lienhart. Univerität Augburg à ÊÇÅÍÆ ËÀǼ Approximating Optimal Viual Senor Placement E. Hörter, R. Lienhart Report 2006-01 Januar 2006 Intitut für Informatik D-86135 Augburg Copyright c E. Hörter, R. Lienhart Intitut

More information

Performance of a Robust Filter-based Approach for Contour Detection in Wireless Sensor Networks

Performance of a Robust Filter-based Approach for Contour Detection in Wireless Sensor Networks Performance of a Robut Filter-baed Approach for Contour Detection in Wirele Senor Network Hadi Alati, William A. Armtrong, Jr., and Ai Naipuri Department of Electrical and Computer Engineering The Univerity

More information

A TOPSIS based Method for Gene Selection for Cancer Classification

A TOPSIS based Method for Gene Selection for Cancer Classification Volume 67 No17, April 2013 A TOPSIS baed Method for Gene Selection for Cancer Claification IMAbd-El Fattah,WIKhedr, KMSallam, 1 Department of Statitic, 3 Department of Deciion upport, 2 Department of information

More information

MAT 155: Describing, Exploring, and Comparing Data Page 1 of NotesCh2-3.doc

MAT 155: Describing, Exploring, and Comparing Data Page 1 of NotesCh2-3.doc MAT 155: Decribing, Exploring, and Comparing Data Page 1 of 8 001-oteCh-3.doc ote for Chapter Summarizing and Graphing Data Chapter 3 Decribing, Exploring, and Comparing Data Frequency Ditribution, Graphic

More information

Key Terms - MinMin, MaxMin, Sufferage, Task Scheduling, Standard Deviation, Load Balancing.

Key Terms - MinMin, MaxMin, Sufferage, Task Scheduling, Standard Deviation, Load Balancing. Volume 3, Iue 11, November 2013 ISSN: 2277 128X International Journal of Advanced Reearch in Computer Science and Software Engineering Reearch Paper Available online at: www.ijarce.com Tak Aignment in

More information

1 The secretary problem

1 The secretary problem Thi i new material: if you ee error, pleae email jtyu at tanford dot edu 1 The ecretary problem We will tart by analyzing the expected runtime of an algorithm, a you will be expected to do on your homework.

More information

SLA Adaptation for Service Overlay Networks

SLA Adaptation for Service Overlay Networks SLA Adaptation for Service Overlay Network Con Tran 1, Zbigniew Dziong 1, and Michal Pióro 2 1 Department of Electrical Engineering, École de Technologie Supérieure, Univerity of Quebec, Montréal, Canada

More information

A note on degenerate and spectrally degenerate graphs

A note on degenerate and spectrally degenerate graphs A note on degenerate and pectrally degenerate graph Noga Alon Abtract A graph G i called pectrally d-degenerate if the larget eigenvalue of each ubgraph of it with maximum degree D i at mot dd. We prove

More information

A METHOD OF REAL-TIME NURBS INTERPOLATION WITH CONFINED CHORD ERROR FOR CNC SYSTEMS

A METHOD OF REAL-TIME NURBS INTERPOLATION WITH CONFINED CHORD ERROR FOR CNC SYSTEMS Vietnam Journal of Science and Technology 55 (5) (017) 650-657 DOI: 10.1565/55-518/55/5/906 A METHOD OF REAL-TIME NURBS INTERPOLATION WITH CONFINED CHORD ERROR FOR CNC SYSTEMS Nguyen Huu Quang *, Banh

More information

CENTER-POINT MODEL OF DEFORMABLE SURFACE

CENTER-POINT MODEL OF DEFORMABLE SURFACE CENTER-POINT MODEL OF DEFORMABLE SURFACE Piotr M. Szczypinki Iintitute of Electronic, Technical Univerity of Lodz, Poland Abtract: Key word: Center-point model of deformable urface for egmentation of 3D

More information

DAROS: Distributed User-Server Assignment And Replication For Online Social Networking Applications

DAROS: Distributed User-Server Assignment And Replication For Online Social Networking Applications DAROS: Ditributed Uer-Server Aignment And Replication For Online Social Networking Application Thuan Duong-Ba School of EECS Oregon State Univerity Corvalli, OR 97330, USA Email: duongba@eec.oregontate.edu

More information

xy-monotone path existence queries in a rectilinear environment

xy-monotone path existence queries in a rectilinear environment CCCG 2012, Charlottetown, P.E.I., Augut 8 10, 2012 xy-monotone path exitence querie in a rectilinear environment Gregory Bint Anil Mahehwari Michiel Smid Abtract Given a planar environment coniting of

More information

Planning of scooping position and approach path for loading operation by wheel loader

Planning of scooping position and approach path for loading operation by wheel loader 22 nd International Sympoium on Automation and Robotic in Contruction ISARC 25 - September 11-14, 25, Ferrara (Italy) 1 Planning of cooping poition and approach path for loading operation by wheel loader

More information

On successive packing approach to multidimensional (M-D) interleaving

On successive packing approach to multidimensional (M-D) interleaving On ucceive packing approach to multidimenional (M-D) interleaving Xi Min Zhang Yun Q. hi ankar Bau Abtract We propoe an interleaving cheme for multidimenional (M-D) interleaving. To achieved by uing a

More information

ANALYSIS OF THE FIRST LAYER IN WEIGHTLESS NEURAL NETWORKS FOR 3_DIMENSIONAL PATTERN RECOGNITION

ANALYSIS OF THE FIRST LAYER IN WEIGHTLESS NEURAL NETWORKS FOR 3_DIMENSIONAL PATTERN RECOGNITION ANALYSIS OF THE FIRST LAYER IN WEIGHTLESS NEURAL NETWORKS FOR 3_DIMENSIONAL PATTERN RECOGNITION A. Váque-Nava * Ecuela de Ingeniería. CENTRO UNIVERSITARIO MEXICO. DIVISION DE ESTUDIOS SUPERIORES J. Figueroa

More information

Hassan Ghaziri AUB, OSB Beirut, Lebanon Key words Competitive self-organizing maps, Meta-heuristics, Vehicle routing problem,

Hassan Ghaziri AUB, OSB Beirut, Lebanon Key words Competitive self-organizing maps, Meta-heuristics, Vehicle routing problem, COMPETITIVE PROBABIISTIC SEF-ORGANIZING MAPS FOR ROUTING PROBEMS Haan Ghaziri AUB, OSB Beirut, ebanon ghaziri@aub.edu.lb Abtract In thi paper, we have applied the concept of the elf-organizing map (SOM)

More information

On combining Learning Vector Quantization and the Bayesian classifiers for natural textured images

On combining Learning Vector Quantization and the Bayesian classifiers for natural textured images On combining Learning Vector Quantization and the Bayeian claifier for natural textured image María Guiarro Dept. Ingeniería del Software e Inteligencia Artificial Facultad Informática Univeridad Complutene

More information

/06/$ IEEE 364

/06/$ IEEE 364 006 IEEE International ympoium on ignal Proceing and Information Technology oie Variance Etimation In ignal Proceing David Makovoz IPAC, California Intitute of Technology, MC-0, Paadena, CA, 95 davidm@ipac.caltech.edu;

More information

Gray-level histogram. Intensity (grey-level) transformation, or mapping. Use of intensity transformations:

Gray-level histogram. Intensity (grey-level) transformation, or mapping. Use of intensity transformations: Faculty of Informatic Eötvö Loránd Univerity Budapet, Hungary Lecture : Intenity Tranformation Image enhancement by point proceing Spatial domain and frequency domain method Baic Algorithm for Digital

More information

Building a Compact On-line MRF Recognizer for Large Character Set using Structured Dictionary Representation and Vector Quantization Technique

Building a Compact On-line MRF Recognizer for Large Character Set using Structured Dictionary Representation and Vector Quantization Technique 202 International Conference on Frontier in Handwriting Recognition Building a Compact On-line MRF Recognizer for Large Character Set uing Structured Dictionary Repreentation and Vector Quantization Technique

More information

(12) Patent Application Publication (10) Pub. No.: US 2011/ A1

(12) Patent Application Publication (10) Pub. No.: US 2011/ A1 (19) United State US 2011 0316690A1 (12) Patent Application Publication (10) Pub. No.: US 2011/0316690 A1 Siegman (43) Pub. Date: Dec. 29, 2011 (54) SYSTEMAND METHOD FOR IDENTIFYING ELECTRICAL EQUIPMENT

More information

Analyzing Hydra Historical Statistics Part 2

Analyzing Hydra Historical Statistics Part 2 Analyzing Hydra Hitorical Statitic Part Fabio Maimo Ottaviani EPV Technologie White paper 5 hnode HSM Hitorical Record The hnode i the hierarchical data torage management node and ha to perform all the

More information

A Multi-objective Genetic Algorithm for Reliability Optimization Problem

A Multi-objective Genetic Algorithm for Reliability Optimization Problem International Journal of Performability Engineering, Vol. 5, No. 3, April 2009, pp. 227-234. RAMS Conultant Printed in India A Multi-objective Genetic Algorithm for Reliability Optimization Problem AMAR

More information

Handling Degenerate Cases in Exact Geodesic Computation on Triangle Meshes

Handling Degenerate Cases in Exact Geodesic Computation on Triangle Meshes The Viual Computer manucript. (will be inerted b the editor) Yong-Jin Liu Qian-Yi Zhou Shi-Min Hu Degenerate Cae in Eact Geodeic Computation on Triangle Mehe Abtract The computation of eact geodeic on

More information

A Linear Interpolation-Based Algorithm for Path Planning and Replanning on Girds *

A Linear Interpolation-Based Algorithm for Path Planning and Replanning on Girds * Advance in Linear Algebra & Matrix Theory, 2012, 2, 20-24 http://dx.doi.org/10.4236/alamt.2012.22003 Publihed Online June 2012 (http://www.scirp.org/journal/alamt) A Linear Interpolation-Baed Algorithm

More information

A Fast Association Rule Algorithm Based On Bitmap and Granular Computing

A Fast Association Rule Algorithm Based On Bitmap and Granular Computing A Fat Aociation Rule Algorithm Baed On Bitmap and Granular Computing T.Y.Lin Xiaohua Hu Eric Louie Dept. of Computer Science College of Information Science IBM Almaden Reearch Center San Joe State Univerity

More information

Increasing Throughput and Reducing Delay in Wireless Sensor Networks Using Interference Alignment

Increasing Throughput and Reducing Delay in Wireless Sensor Networks Using Interference Alignment Int. J. Communication, Network and Sytem Science, 0, 5, 90-97 http://dx.doi.org/0.436/ijcn.0.50 Publihed Online February 0 (http://www.scirp.org/journal/ijcn) Increaing Throughput and Reducing Delay in

More information

Keywords: Defect detection, linear phased array transducer, parameter optimization, phased array ultrasonic B-mode imaging testing.

Keywords: Defect detection, linear phased array transducer, parameter optimization, phased array ultrasonic B-mode imaging testing. Send Order for Reprint to reprint@benthamcience.ae 488 The Open Automation and Control Sytem Journal, 2014, 6, 488-492 Open Acce Parameter Optimization of Linear Phaed Array Tranducer for Defect Detection

More information

Laboratory Exercise 6

Laboratory Exercise 6 Laboratory Exercie 6 Adder, Subtractor, and Multiplier The purpoe of thi exercie i to examine arithmetic circuit that add, ubtract, and multiply number. Each type of circuit will be implemented in two

More information

ISSN: (Online) Volume 3, Issue 4, April 2015 International Journal of Advance Research in Computer Science and Management Studies

ISSN: (Online) Volume 3, Issue 4, April 2015 International Journal of Advance Research in Computer Science and Management Studies ISSN: 2321-7782 (Online) Volume 3, Iue 4, April 2015 International Journal Advance Reearch in Computer Science and Management Studie Reearch Article / Survey Paper / Cae Study Available online at: www.ijarcm.com

More information

Texture-Constrained Active Shape Models

Texture-Constrained Active Shape Models 107 Texture-Contrained Active Shape Model Shuicheng Yan, Ce Liu Stan Z. Li Hongjiang Zhang Heung-Yeung Shum Qianheng Cheng Microoft Reearch Aia, Beijing Sigma Center, Beijing 100080, China Dept. of Info.

More information

Karen L. Collins. Wesleyan University. Middletown, CT and. Mark Hovey MIT. Cambridge, MA Abstract

Karen L. Collins. Wesleyan University. Middletown, CT and. Mark Hovey MIT. Cambridge, MA Abstract Mot Graph are Edge-Cordial Karen L. Collin Dept. of Mathematic Weleyan Univerity Middletown, CT 6457 and Mark Hovey Dept. of Mathematic MIT Cambridge, MA 239 Abtract We extend the definition of edge-cordial

More information

Shortest Paths Problem. CS 362, Lecture 20. Today s Outline. Negative Weights

Shortest Paths Problem. CS 362, Lecture 20. Today s Outline. Negative Weights Shortet Path Problem CS 6, Lecture Jared Saia Univerity of New Mexico Another intereting problem for graph i that of finding hortet path Aume we are given a weighted directed graph G = (V, E) with two

More information

IMPROVED JPEG DECOMPRESSION OF DOCUMENT IMAGES BASED ON IMAGE SEGMENTATION. Tak-Shing Wong, Charles A. Bouman, and Ilya Pollak

IMPROVED JPEG DECOMPRESSION OF DOCUMENT IMAGES BASED ON IMAGE SEGMENTATION. Tak-Shing Wong, Charles A. Bouman, and Ilya Pollak IMPROVED DECOMPRESSION OF DOCUMENT IMAGES BASED ON IMAGE SEGMENTATION Tak-Shing Wong, Charle A. Bouman, and Ilya Pollak School of Electrical and Computer Engineering Purdue Univerity ABSTRACT We propoe

More information

A reduced reference image quality metric based on feature fusion and neural networks

A reduced reference image quality metric based on feature fusion and neural networks Univerity of Wollongong Reearch Online Faculty of Engineering and Information Science - Paper: Part A Faculty of Engineering and Information Science 2011 A reduced reference image quality metric baed on

More information

Touring a Sequence of Polygons

Touring a Sequence of Polygons Touring a Sequence of Polygon Mohe Dror (1) Alon Efrat (1) Anna Lubiw (2) Joe Mitchell (3) (1) Univerity of Arizona (2) Univerity of Waterloo (3) Stony Brook Univerity Problem: Given a equence of k polygon

More information

Shortest Paths with Single-Point Visibility Constraint

Shortest Paths with Single-Point Visibility Constraint Shortet Path with Single-Point Viibility Contraint Ramtin Khoravi Mohammad Ghodi Department of Computer Engineering Sharif Univerity of Technology Abtract Thi paper tudie the problem of finding a hortet

More information

Shortest Path Routing in Arbitrary Networks

Shortest Path Routing in Arbitrary Networks Journal of Algorithm, Vol 31(1), 1999 Shortet Path Routing in Arbitrary Network Friedhelm Meyer auf der Heide and Berthold Vöcking Department of Mathematic and Computer Science and Heinz Nixdorf Intitute,

More information

Touring a Sequence of Polygons

Touring a Sequence of Polygons Touring a Sequence of Polygon Mohe Dror (1) Alon Efrat (1) Anna Lubiw (2) Joe Mitchell (3) (1) Univerity of Arizona (2) Univerity of Waterloo (3) Stony Brook Univerity Problem: Given a equence of k polygon

More information

New Structural Decomposition Techniques for Constraint Satisfaction Problems

New Structural Decomposition Techniques for Constraint Satisfaction Problems New Structural Decompoition Technique for Contraint Satifaction Problem Yaling Zheng and Berthe Y. Choueiry Contraint Sytem Laboratory Univerity of Nebraka-Lincoln Email: yzheng choueiry@ce.unl.edu Abtract.

More information

Stochastic Search and Graph Techniques for MCM Path Planning Christine D. Piatko, Christopher P. Diehl, Paul McNamee, Cheryl Resch and I-Jeng Wang

Stochastic Search and Graph Techniques for MCM Path Planning Christine D. Piatko, Christopher P. Diehl, Paul McNamee, Cheryl Resch and I-Jeng Wang Stochatic Search and Graph Technique for MCM Path Planning Chritine D. Piatko, Chritopher P. Diehl, Paul McNamee, Cheryl Rech and I-Jeng Wang The John Hopkin Univerity Applied Phyic Laboratory, Laurel,

More information

Today s Outline. CS 561, Lecture 23. Negative Weights. Shortest Paths Problem. The presence of a negative cycle might mean that there is

Today s Outline. CS 561, Lecture 23. Negative Weights. Shortest Paths Problem. The presence of a negative cycle might mean that there is Today Outline CS 56, Lecture Jared Saia Univerity of New Mexico The path that can be trodden i not the enduring and unchanging Path. The name that can be named i not the enduring and unchanging Name. -

More information

How to Select Measurement Points in Access Point Localization

How to Select Measurement Points in Access Point Localization Proceeding of the International MultiConference of Engineer and Computer Scientit 205 Vol II, IMECS 205, March 8-20, 205, Hong Kong How to Select Meaurement Point in Acce Point Localization Xiaoling Yang,

More information

A DIVISIVE HIERARCHICAL CLUSTERING- BASED METHOD FOR INDEXING IMAGE INFORMATION

A DIVISIVE HIERARCHICAL CLUSTERING- BASED METHOD FOR INDEXING IMAGE INFORMATION A DIVISIVE HIERARCHICAL CLUSTERING- BASED METHOD FOR INDEXING IMAGE INFORMATION ABSTRACT Najva Izadpanah Department of Computer Engineering, Ilamic Azad Univerity, Qazvin Branch, Qazvin, Iran In mot practical

More information

Multiclass Road Sign Detection using Multiplicative Kernel

Multiclass Road Sign Detection using Multiplicative Kernel Proceeding of the Croatian Computer Viion Workhop, Year 1 Multicla Road Sign Detection uing Multiplicative Kernel Valentina Zadrija Mireo d. d. Zagreb, Croatia valentina.zadrija@mireo.hr Siniša Šegvić

More information

Floating Point CORDIC Based Power Operation

Floating Point CORDIC Based Power Operation Floating Point CORDIC Baed Power Operation Kazumi Malhan, Padmaja AVL Electrical and Computer Engineering Department School of Engineering and Computer Science Oakland Univerity, Rocheter, MI e-mail: kmalhan@oakland.edu,

More information

CSE 250B Assignment 4 Report

CSE 250B Assignment 4 Report CSE 250B Aignment 4 Report March 24, 2012 Yuncong Chen yuncong@c.ucd.edu Pengfei Chen pec008@ucd.edu Yang Liu yal060@c.ucd.edu Abtract In thi project, we implemented the recurive autoencoder (RAE) a decribed

More information

Restricted Nearest Feature Line with Ellipse for Face Recognition

Restricted Nearest Feature Line with Ellipse for Face Recognition Journal of Information Hiding and Multimedia Signal Processing c 2012 ISSN 2073-4212 Ubiquitous International Volume 3, Number 3, July 2012 Restricted Nearest Feature Line with Ellipse for Face Recognition

More information

Stress-Blended Eddy Simulation (SBES) - A new Paradigm in hybrid RANS-LES Modeling

Stress-Blended Eddy Simulation (SBES) - A new Paradigm in hybrid RANS-LES Modeling Stre-Blended Eddy Simulation (SBES) - A new Paradigm in hybrid RANS-LES Modeling Menter F.R. ANSYS Germany GmbH Introduction It i oberved in many CFD imulation that RANS model how inherent technology limitation

More information

UC Berkeley International Conference on GIScience Short Paper Proceedings

UC Berkeley International Conference on GIScience Short Paper Proceedings UC Berkeley International Conference on GIScience Short Paper Proceeding Title A novel method for probabilitic coverage etimation of enor network baed on 3D vector repreentation in complex urban environment

More information

Multicast with Network Coding in Application-Layer Overlay Networks

Multicast with Network Coding in Application-Layer Overlay Networks IEEE JOURNAL ON SELECTED AREAS IN COMMUNICATIONS, VOL. 22, NO. 1, JANUARY 2004 1 Multicat with Network Coding in Application-Layer Overlay Network Ying Zhu, Baochun Li, Member, IEEE, and Jiang Guo Abtract

More information

Refining SIRAP with a Dedicated Resource Ceiling for Self-Blocking

Refining SIRAP with a Dedicated Resource Ceiling for Self-Blocking Refining SIRAP with a Dedicated Reource Ceiling for Self-Blocking Mori Behnam, Thoma Nolte Mälardalen Real-Time Reearch Centre P.O. Box 883, SE-721 23 Väterå, Sweden {mori.behnam,thoma.nolte}@mdh.e ABSTRACT

More information

Delaunay Triangulation: Incremental Construction

Delaunay Triangulation: Incremental Construction Chapter 6 Delaunay Triangulation: Incremental Contruction In the lat lecture, we have learned about the Lawon ip algorithm that compute a Delaunay triangulation of a given n-point et P R 2 with O(n 2 )

More information

Keywords Cloud Computing, Service Level Agreements (SLA), CloudSim, Monitoring & Controlling SLA Agent, JADE

Keywords Cloud Computing, Service Level Agreements (SLA), CloudSim, Monitoring & Controlling SLA Agent, JADE Volume 5, Iue 8, Augut 2015 ISSN: 2277 128X International Journal of Advanced Reearch in Computer Science and Software Engineering Reearch Paper Available online at: www.ijarce.com Verification of Agent

More information

Advanced Encryption Standard and Modes of Operation

Advanced Encryption Standard and Modes of Operation Advanced Encryption Standard and Mode of Operation G. Bertoni L. Breveglieri Foundation of Cryptography - AES pp. 1 / 50 AES Advanced Encryption Standard (AES) i a ymmetric cryptographic algorithm AES

More information

TAM 212 Worksheet 3. Solutions

TAM 212 Worksheet 3. Solutions Name: Group member: TAM 212 Workheet 3 Solution The workheet i concerned with the deign of the loop-the-loop for a roller coater ytem. Old loop deign: The firt generation of loop wa circular, a hown below.

More information

Maneuverable Relays to Improve Energy Efficiency in Sensor Networks

Maneuverable Relays to Improve Energy Efficiency in Sensor Networks Maneuverable Relay to Improve Energy Efficiency in Senor Network Stephan Eidenbenz, Luka Kroc, Jame P. Smith CCS-5, MS M997; Lo Alamo National Laboratory; Lo Alamo, NM 87545. Email: {eidenben, kroc, jpmith}@lanl.gov

More information

IMPLEMENTATION OF AREA, VOLUME AND LINE SOURCES

IMPLEMENTATION OF AREA, VOLUME AND LINE SOURCES December 01 ADMS 5 P503I1 IMPEMENTATION OF AREA, VOUME AND INE SOURCES The Met. Office (D J Thomon) and CERC 1. INTRODUCTION ADMS model line ource, and area and volume ource with conve polgon bae area.

More information

Topics. Lecture 37: Global Optimization. Issues. A Simple Example: Copy Propagation X := 3 B > 0 Y := 0 X := 4 Y := Z + W A := 2 * 3X

Topics. Lecture 37: Global Optimization. Issues. A Simple Example: Copy Propagation X := 3 B > 0 Y := 0 X := 4 Y := Z + W A := 2 * 3X Lecture 37: Global Optimization [Adapted from note by R. Bodik and G. Necula] Topic Global optimization refer to program optimization that encompa multiple baic block in a function. (I have ued the term

More information

KS3 Maths Assessment Objectives

KS3 Maths Assessment Objectives KS3 Math Aement Objective Tranition Stage 9 Ratio & Proportion Probabilit y & Statitic Appreciate the infinite nature of the et of integer, real and rational number Can interpret fraction and percentage

More information

Data Mining with Linguistic Thresholds

Data Mining with Linguistic Thresholds Int. J. Contemp. Math. Science, Vol. 7, 22, no. 35, 7-725 Data Mining with Linguitic Threhold Tzung-Pei Hong Department of Electrical Engineering National Univerity of Kaohiung Kaohiung, Taiwan, R.O.C.

More information

[N309] Feedforward Active Noise Control Systems with Online Secondary Path Modeling. Muhammad Tahir Akhtar, Masahide Abe, and Masayuki Kawamata

[N309] Feedforward Active Noise Control Systems with Online Secondary Path Modeling. Muhammad Tahir Akhtar, Masahide Abe, and Masayuki Kawamata he 32nd International Congre and Expoition on Noie Control Engineering Jeju International Convention Center, Seogwipo, Korea, Augut 25-28, 2003 [N309] Feedforward Active Noie Control Sytem with Online

More information

Localized Minimum Spanning Tree Based Multicast Routing with Energy-Efficient Guaranteed Delivery in Ad Hoc and Sensor Networks

Localized Minimum Spanning Tree Based Multicast Routing with Energy-Efficient Guaranteed Delivery in Ad Hoc and Sensor Networks Localized Minimum Spanning Tree Baed Multicat Routing with Energy-Efficient Guaranteed Delivery in Ad Hoc and Senor Network Hanne Frey Univerity of Paderborn D-3398 Paderborn hanne.frey@uni-paderborn.de

More information

Distributed Partial Information Management (DPIM) Schemes for Survivable Networks - Part II

Distributed Partial Information Management (DPIM) Schemes for Survivable Networks - Part II IEEE INFOCO 2002 1 Ditributed Partial Information anagement (DPI) Scheme for Survivable Network - Part II Dahai Xu Chunming Qiao Department of Computer Science and Engineering State Univerity of New York

More information

Using Mouse Feedback in Computer Assisted Transcription of Handwritten Text Images

Using Mouse Feedback in Computer Assisted Transcription of Handwritten Text Images 2009 10th International Conference on Document Analyi and Recognition Uing Moue Feedback in Computer Aited Trancription of Handwritten Text Image Verónica Romero, Alejandro H. Toelli, Enrique Vidal Intituto

More information

Numerical Modeling of Material Discontinuity Using Mixed MLPG Collocation Method

Numerical Modeling of Material Discontinuity Using Mixed MLPG Collocation Method umerical odeling of aterial Dicontinuit Uing ied LPG Collocation ethod B. alušić 1,. Sorić 1 and T. arak 1 Abtract A mied LPG collocation method i applied for the modeling of material dicontinuit in heterogeneou

More information

Laboratory Exercise 2

Laboratory Exercise 2 Laoratory Exercie Numer and Diplay Thi i an exercie in deigning cominational circuit that can perform inary-to-decimal numer converion and inary-coded-decimal (BCD) addition. Part I We wih to diplay on

More information

USING ARTIFICIAL NEURAL NETWORKS TO APPROXIMATE A DISCRETE EVENT STOCHASTIC SIMULATION MODEL

USING ARTIFICIAL NEURAL NETWORKS TO APPROXIMATE A DISCRETE EVENT STOCHASTIC SIMULATION MODEL USING ARTIFICIAL NEURAL NETWORKS TO APPROXIMATE A DISCRETE EVENT STOCHASTIC SIMULATION MODEL Robert A. Kilmer Department of Sytem Engineering Unite State Military Acaemy Wet Point, NY 1996 Alice E. Smith

More information

Representations and Transformations. Objectives

Representations and Transformations. Objectives Repreentation and Tranformation Objective Derive homogeneou coordinate tranformation matrice Introduce tandard tranformation - Rotation - Tranlation - Scaling - Shear Scalar, Point, Vector Three baic element

More information

A Practical Model for Minimizing Waiting Time in a Transit Network

A Practical Model for Minimizing Waiting Time in a Transit Network A Practical Model for Minimizing Waiting Time in a Tranit Network Leila Dianat, MASc, Department of Civil Engineering, Sharif Univerity of Technology, Tehran, Iran Youef Shafahi, Ph.D. Aociate Profeor,

More information

3D MODELLING WITH LINEAR APPROACHES USING GEOMETRIC PRIMITIVES

3D MODELLING WITH LINEAR APPROACHES USING GEOMETRIC PRIMITIVES MAKARA, TEKNOLOGI, VOL. 9, NO., APRIL 5: 3-35 3D MODELLING WITH LINEAR APPROACHES USING GEOMETRIC PRIMITIVES Mochammad Zulianyah Informatic Engineering, Faculty of Engineering, ARS International Univerity,

More information

A PROBABILISTIC NOTION OF CAMERA GEOMETRY: CALIBRATED VS. UNCALIBRATED

A PROBABILISTIC NOTION OF CAMERA GEOMETRY: CALIBRATED VS. UNCALIBRATED A PROBABILISTIC NOTION OF CAMERA GEOMETRY: CALIBRATED VS. UNCALIBRATED Jutin Domke and Yianni Aloimono Computational Viion Laboratory, Center for Automation Reearch Univerity of Maryland College Park,

More information

Evolution of Non-Deterministic Incremental Algorithms. Hugues Juille. Volen Center for Complex Systems. Brandeis University. Waltham, MA

Evolution of Non-Deterministic Incremental Algorithms. Hugues Juille. Volen Center for Complex Systems. Brandeis University. Waltham, MA Evolution of Non-Determinitic Incremental Algorithm a a New Approach for Search in State Space Hugue Juille Computer Science Department Volen Center for Complex Sytem Brandei Univerity Waltham, MA 02254-9110

More information

An Approach to a Test Oracle for XML Query Testing

An Approach to a Test Oracle for XML Query Testing An Approach to a Tet Oracle for XML Query Teting Dae S. Kim-Park, Claudio de la Riva, Javier Tuya Univerity of Oviedo Computing Department Campu of Vieque, /n, 33204 (SPAIN) kim_park@li.uniovi.e, claudio@uniovi.e,

More information

PARALLEL knn ON GPU ARCHITECTURE USING OpenCL

PARALLEL knn ON GPU ARCHITECTURE USING OpenCL PARALLEL knn ON GPU ARCHITECTURE USING OpenCL V.B.Nikam 1, B.B.Mehram 2 1 Aociate Profeor, Department of Computer Engineering and Information Technology, Jijabai Technological Intitute, Matunga, Mumbai,

More information

Image authentication and tamper detection using fragile watermarking in spatial domain

Image authentication and tamper detection using fragile watermarking in spatial domain International Journal of Advanced Reearch in Computer Engineering & Technology (IJARCET) Volume 6, Iue 7, July 2017, ISSN: 2278 1323 Image authentication and tamper detection uing fragile watermarking

More information

arxiv: v1 [cs.ds] 27 Feb 2018

arxiv: v1 [cs.ds] 27 Feb 2018 Incremental Strong Connectivity and 2-Connectivity in Directed Graph Louka Georgiadi 1, Giueppe F. Italiano 2, and Niko Parotidi 2 arxiv:1802.10189v1 [c.ds] 27 Feb 2018 1 Univerity of Ioannina, Greece.

More information

Distributed Packet Processing Architecture with Reconfigurable Hardware Accelerators for 100Gbps Forwarding Performance on Virtualized Edge Router

Distributed Packet Processing Architecture with Reconfigurable Hardware Accelerators for 100Gbps Forwarding Performance on Virtualized Edge Router Ditributed Packet Proceing Architecture with Reconfigurable Hardware Accelerator for 100Gbp Forwarding Performance on Virtualized Edge Router Satohi Nihiyama, Hitohi Kaneko, and Ichiro Kudo Abtract To

More information

See chapter 8 in the textbook. Dr Muhammad Al Salamah, Industrial Engineering, KFUPM

See chapter 8 in the textbook. Dr Muhammad Al Salamah, Industrial Engineering, KFUPM Goal programming Objective of the topic: Indentify indutrial baed ituation where two or more objective function are required. Write a multi objective function model dla a goal LP Ue weighting um and preemptive

More information

ALTHOUGH deep neural networks trained on large-scale

ALTHOUGH deep neural networks trained on large-scale 1 Unupervied Domain Adaptation with Adverarial Reidual Tranform Network Guanyu Cai, Yuqin Wang, Mengchu Zhou, Fellow, IEEE, Lianghua He arxiv:1804.09578v1 [c.cv] 25 Apr 2018 Abtract Domain adaptation i

More information

An Algebraic Approach to Adaptive Scalable Overlay Network Monitoring

An Algebraic Approach to Adaptive Scalable Overlay Network Monitoring An Algebraic Approach to Adaptive Scalable Overlay Network Monitoring ABSTRACT Overlay network monitoring enable ditributed Internet application to detect and recover from path outage and period of degraded

More information

Drawing Lines in 2 Dimensions

Drawing Lines in 2 Dimensions Drawing Line in 2 Dimenion Drawing a traight line (or an arc) between two end point when one i limited to dicrete pixel require a bit of thought. Conider the following line uperimpoed on a 2 dimenional

More information

CERIAS Tech Report EFFICIENT PARALLEL ALGORITHMS FOR PLANAR st-graphs. by Mikhail J. Atallah, Danny Z. Chen, and Ovidiu Daescu

CERIAS Tech Report EFFICIENT PARALLEL ALGORITHMS FOR PLANAR st-graphs. by Mikhail J. Atallah, Danny Z. Chen, and Ovidiu Daescu CERIAS Tech Report 2003-15 EFFICIENT PARALLEL ALGORITHMS FOR PLANAR t-graphs by Mikhail J. Atallah, Danny Z. Chen, and Ovidiu Daecu Center for Education and Reearch in Information Aurance and Security,

More information

Performance Evaluation of an Advanced Local Search Evolutionary Algorithm

Performance Evaluation of an Advanced Local Search Evolutionary Algorithm Anne Auger and Nikolau Hanen Performance Evaluation of an Advanced Local Search Evolutionary Algorithm Proceeding of the IEEE Congre on Evolutionary Computation, CEC 2005 c IEEE Performance Evaluation

More information

Research on Star Image Noise Filtering Based on Diffusion Model of Regularization Influence Function

Research on Star Image Noise Filtering Based on Diffusion Model of Regularization Influence Function 016 Sith International Conference on Intrumentation & Meaurement Computer Communication and Control Reearch on Star Image Noie Filtering Baed on Diffuion Model of Regularization Influence Function SunJianming

More information

Multi-Target Tracking In Clutter

Multi-Target Tracking In Clutter Multi-Target Tracking In Clutter John N. Sander-Reed, Mary Jo Duncan, W.B. Boucher, W. Michael Dimmler, Shawn O Keefe ABSTRACT A high frame rate (0 Hz), multi-target, video tracker ha been developed and

More information

Comparison of Methods for Horizon Line Detection in Sea Images

Comparison of Methods for Horizon Line Detection in Sea Images Comparion of Method for Horizon Line Detection in Sea Image Tzvika Libe Evgeny Gerhikov and Samuel Koolapov Department of Electrical Engineering Braude Academic College of Engineering Karmiel 2982 Irael

More information

On the Use of Shadows in Stance Recovery

On the Use of Shadows in Stance Recovery On the Ue of Shadow in Stance Recovery Alfred M. Brucktein, 1 Robert J. Holt, 1 Yve D. Jean, 2 Arun N. Netravali 1 1 Bell Laboratorie, Lucent Technologie, Murray Hill, NJ 094 2 Avaya Communication, Murray

More information

Web Science and additionality

Web Science and additionality Admin tuff... Lecture 1: EITN01 Web Intelligence and Information Retrieval Meage, lide, handout, lab manual and link: http://www.eit.lth.e/coure/eitn01 Contact: Ander Ardö, Ander.Ardo@eit.lth.e, room:

More information

SECTOR BASED MULTICAST ROUTING ALGORITHM FOR MOBILE AD-HOC NETWORKS

SECTOR BASED MULTICAST ROUTING ALGORITHM FOR MOBILE AD-HOC NETWORKS SECTOR BASED MULTICAST ROUTING ALGORITHM OR MOBILE AD-HOC NETWORKS Murali Paramewaran 1 and Chittaranjan Hota 2 1 Department of Computer Science & Information Sytem, BITS-Pilani, Pilani, India 2 Department

More information

New Structural Decomposition Techniques for Constraint Satisfaction Problems

New Structural Decomposition Techniques for Constraint Satisfaction Problems 113 New Structural Decompoition Technique for Contraint Satifaction Problem Yaling Zheng and Berthe Y. Choueiry Contraint Sytem Laboratory, Univerity of Nebraka-Lincoln {yzheng,choueiry}@ce.unl.edu Abtract.

More information

A SIMPLE IMPERATIVE LANGUAGE THE STORE FUNCTION NON-TERMINATING COMMANDS

A SIMPLE IMPERATIVE LANGUAGE THE STORE FUNCTION NON-TERMINATING COMMANDS A SIMPLE IMPERATIVE LANGUAGE Eventually we will preent the emantic of a full-blown language, with declaration, type and looping. However, there are many complication, o we will build up lowly. Our firt

More information

Cut Tree Construction from Massive Graphs

Cut Tree Construction from Massive Graphs Cut Tree Contruction from Maive Graph Takuya Akiba Yoichi Iwata Youke Samehima # Naoto Mizuno Youke Yano # Preferred Network, Inc. National Intitute of Informatic # Recruit Holding Co., Ltd. The Univerity

More information

A New Approach for In-Vehicle Camera Traffic Sign Detection and Recognition.

A New Approach for In-Vehicle Camera Traffic Sign Detection and Recognition. A New Approach for In-Vehicle Camera Traffic Sign Detection and Recognition. Andrzej Ruta, Yongmin Li School of Information Sytem, Computing & Mathematic Brunel Univerity Uxbridge, Middleex UB8 3PH, United

More information