Fast CU Partition Strategy for HEVC Intra-Frame Coding Using Learning Approach via Random Forests

Size: px
Start display at page:

Download "Fast CU Partition Strategy for HEVC Intra-Frame Coding Using Learning Approach via Random Forests"

Transcription

1 Fast CU Partton Strategy for HEVC Intra-Frame Codng Usng Learnng Approach va Random Forests Bochuan Du, Wan-Ch Su and Xuefe Yang Centre of multmeda sgnal processng, Department of Electronc and Informaton Engneerng The Hong Kong Polytechnc Unversty, Hong Kong E-mal: Tel: Abstract HEVC (Hgh Effcency Vdeo Codng) acheves cuttng edge encodng effcency and outperforms prevous standards, such as the H.264/AVC. One of the key contrbutons to the mprovement s the ntra-frame codng that employs abundant codng unt (CU) szes. However fndng the optmal CU sze s computatonally expensve. To allevate the ntra encodng complexty and facltate the real-tme mplementaton, we use a machne learnng technque: the random forests, for tranng. Based on off-lne tranng, we propose usng the forest classfer to skp or termnate the current CU depth level. In addton, neghborng CU sze decsons are utlzed to determne the current depth range. Expermental results show that our proposed algorthm can acheve 48.31% tme reducton, wth 0.80% ncrease n the Bjøntegaard delta btrate (), whch are state-of-the-art results compared wth all algorthms n the lterature. I. INTRODUCTION In 2013, the Jont Collaboratve Team on Vdeo Codng (JCT-VC) has come up wth a mlestone of vdeo compresson standard, known as Hgh Effcency Vdeo Codng (HEVC) [1], whch s an evolved verson of H.264 [2]. HEVC outperforms H.264 by ncreasng the compresson rate by about 50% and obtanng a hgh encodng effcency for resoluton beyond Hgh Defnton (HD) vdeos, such as the latest 4K and 8K formats vdeos[3][4]. Compared wth H.264, HEVC has two man enhancements n encoder, ) nested quad-tree structures wth codng unt (CU), predcton unt (PU), and transform unt (TU) [5] and ) numerous predcton modes [4]. One of the consderable areas n HEVC s ntra codng, for whch blocks of pxels are coded only wth references to the current frame. It explots the spatal correlaton on blocks wth 35 predcton modes and CU szes varyng from block sze 64x64 down to 8x8 [6]. These new features affect the complexty extremely, makng t dffcult for real-tme mplementaton, especally for HD or Ultra HD vdeos that have hgh resolutons. The fact motvates researches on early decson of CU sze to reduce the codng complexty and acheve a fast encodng speed. Authors n [7] proposed an onlne updated statstcal parameter to set the threshold for early decson on CU sze splttng and prunng. In [8], texture measurements on each CU block are utlzed to bypass and termnate CU sze decson. Edge complexty n several drectons are used to decde the parttonng of a CU n [10]. However, most of these exstng algorthms manly apply conventonal contentbased or statstc-based methods. Lmted number of papers n the lterature consder to used learnng-based algorthms, whch s currently a hotspot and wdely appled n recognton, classfcaton and sgnal processng. One of the latest lterature that uses learnng-based method on CU sze decson n ntra frame predcton s [9], where data mnng classfers are used. But the tme savng rato mght be far from the stateof-the-art results and the testng sequences used n the paper appear ncomplete for a full evaluaton. Many papers n lterature that explore fast ntra codng algorthms va predcton mode selecton [11][12], whch s also a hot topc for ntra predcton research, whle solely applyng fast predcton mode selecton usually cannot gve an effectve complexty reducton n HEVC. In ths paper, we propose an effcent machne learnng algorthm based on fast CU depth decson, where we manly apply random forests for the learnng. For the rest of ths paper, we wll gve a bref ntroducton of HEVC ntra-frame predcton n secton II, manly focusng on CU sze decson. Secton III covers the detals of our proposed algorthm, whch ncludes CU range decson and the fast splt or nonsplt strategy. In secton IV, we wll show and compare expermental results. Fnally, the concluson and references are provded. II. INTRA PREDICTION IN HEVC Intra predcton s a block-based predcton. Compared wth the H.264 standard where three block szes wth four (for 16x16 block sze) and nne (for 8x8 and 4x4 block sze) predcton modes are avalable to predct a block wthn the current frame [1], the HEVC not only enlarges the block szes to 64x64, 32x32, 16x16, 8x8 and 4x4, but also ncreases the set of predcton modes to 35 [2]. HEVC starts the predcton from a Codng Tree Unt (CTU) whose sze s 64x64. It wll then be further splt nto four 32x32 CUs, each of the four sub-cus can also be splt teratvely nto a smaller quad-tree-based structure [6]. Fgure 1 gves an example of the CTU sze decson on both the tree structure and correspondng block parttons. Depth 0 refers to the block sze 64x64. And the number 0 and 1 represent the APSIPA 1085 APSIPA ASC 2015

2 non-splt and splt decson. If the cost of predcton on the four 32x32 blocks s lower than the cost of predctng 64x64 block, the depth level wll be ncreased by one and goes nto the analyss of each 32x32 blocks. Smlarly, a 32x32 CU wll be compared wth ts four 16x16 sub-cus. When t comes to 8x8 CU, t may be dvded nto PUs wth sze of 4x4. Fnally, the parttonng ends untl the mnmum cost n each level s found. and we have M=4, ncludng the left CU, left-above CU, above CU and rght-above CU. w s the weghtng factor determned by the correlaton between the current CU and each neghborng CU. The sum of these four weghts s normalzed to 1. Due to a stronger correlaton, we assgn w =0.3 for the left CU and above CU. For the left-above CU and rght-above CU whch have weaker correlaton wth the current CU, we let w =0.2. Ths arrangement s for the sake of smplcty. After obtanng the Depth pred the current CU depth range s assgned as below. (2) Fg. 1 An example of CU parttons In each CU, 35 predcton modes, ncludng 2 modes for smooth texture predcton and 33 angular predcton modes, are compared by Rough Mode Decson (RMD) [13][14][15] and Rate Dstorton Optmzaton (RDO) [6]. Then the optmum mode wth the lowest cost s obtaned. Due to numerous amount of possbltes to be evaluated, HEVC s tme-consumng. Ths fact evokes our approach to realze an early CU sze decson wthout gong through the tree structure. III. OUR PROPOSED FAST CU DECISION A. CU Depth Range Decson In a natural scene, neghborng blocks have a strong correlaton wth the current CU. It s known that the ntraframe codng makes use of the encoded above, left, left-above and rght-above CUs as shown n Fgure 2 to reduce computatonal burden n the current CU. Fg.2 Current CU and ts Neghbor Depth Decson These neghborng CUs usually have smlar contents, but HEVC yet has no good methods to ad the current CU sze decson by utlzng ts neghbors. We can make use of the spatal correlaton of them, and predct a depth level as below. M 1 0 Depth w Depth (1) pred where Depth s the depth level of the neghborng CUs, rangng from 0 to 3. M s the number of neghborng blocks, When the predcted depth s smaller than 1, t means the neghborng blocks are flat and homogeneous. The current depth range can be set from 0 to 2, where the smallest CU sze for predcton s lmted to 16x16 (depth 2). When the predcted level s larger than 2, ths reveals that the neghborng CUs are complex and nhomogeneous. We wll set the predcton range from 1 to 3, whch means the current CU predcton starts from block sze 32x32 (depth 1), and searches towards 8x8 (depth 3). For the rest cases of the predcted depth, the predcton range remans unchanged, whch s the same as the default HEVC settngs. Ths approach s smlar to reference [8] and [16], except that [8] s lmted to bypass the current block sze when the current depth level s smaller than the predcted depth mnus one. And [16] defnes the depth range by consderng the ranges of the left and above CU only. B. Early CU Termnaton and Bypass Decson by Random Forests In ths secton, we ntroduce the core of our proposed algorthm. Frst, our forest desgn s gven wth a bref ntroducton of decson tree and random forests. Then a splt algorthm s ntroduced for the optmum splt decson. Lastly, we wll provde detaled strateges for the tranng, and the evaluatons of the tranng results. 1) Forest Desgn The approach random forests [17]-[20], T={T t}, s a combnaton of a set of decson trees {T t} whch jontly realze classfcaton. Each tree has three types of nodes: A root node that has no ncomng edges wth zero or more outgong edges. Internal nodes, each of whch has exactly one ncomng edge and two or more outgong edges. Leaf or Termnal nodes, each of whch has exactly one ncomng edge and no outgong edges. The predcton starts wth the root node. For each non-leaf node, whch ncludes the root and nternal nodes, the procedure allows an ncomng mage block to go to the left or to the rght based on the value of a certan varable whose ndex s stored n the current node. The tree s bult recursvely, at each node the recursve procedure may stop when: APSIPA 1086 APSIPA ASC 2015

3 Depth of the tree branch has reached the maxmum defned value. Number of tranng samples s less than a specfed threshold when further splt s not statstcally representatve. All the samples belong to the same class. In our desgned forests, all the trees are traned wth dfferent tranng sets but wth the same set of parameters. These sets are generated from the orgnal tranng set usng the bootstrap procedure. We totally uses 10 trees, where each of them has a maxmum depths of 25. We apply the luma pxel ntensty n each block as the nput feature vector, and the output classfcaton s 0 for non-splt decson and 1 for splt decson. As shown n Fgure 3, the classfcaton works as below: each random tree classfer takes the nput mage block, classfy t by gong through the decson tree, and outputs the class label 1 or 0 for each tree. The fnal decson s made at the classfcaton result that receved the majorty of votes from all the 10 outputs. Fg. 3 Random forest wth a set of decson trees. 2) Splt Algorthm In the tranng of each non-leaf node, a set of thresholds s tested. If the pxel value s less than the threshold, the mage block goes to the left, otherwse t goes to the rght. The optmum threshold can be obtaned by many measures. In our experments, we use the Gn mpurty crteron [19] c 1 Gn ( t) 1 [ p( t)] (3) 0 where c s the number of classes (here we have c=2), p( t) denotes the rato of data belongng to class at the node t. If the Gn mpurty s lower, the classfcaton performance s better. Then, by gong through all possble thresholds, the optmum one s obtaned wth the lowest Gn mpurty. For example, there are two thresholds Th 1, Th 2 need to be tested. Threshold Th 1 results n 10 tranng samples gong to the left node, where 8 of them are classfed as 0, and the rest 2 of them are classfed as 1. The Gn mpurty for Th 1 s Gn Th1 1 ( ) ( ) 0.32 (4) Threshold Th 2 results n 20 tranng samples gong to the left node, where 10 of them are classfed as 0, the other 10 samples are classfed as 1. The Gn mpurty for Th 2 s Gn Th2 1 ( ) ( ) 0.50 (5) Snce Gn Th1 < Gn Th2, Threshold Th 1 s then selected as the optmum threshold for the splt. 3) Tranng Scheme and Evaluaton Feature selecton ams to select attrbutes that can realze a hgh classfcaton performance. In our proposed method, we take the luma pxel ntensty as the nput attrbutes for the CU sze 64x64 and 32x32 determnaton. The classfcaton result s set to 0 whch means to stay n the current depth wthout any further splt and 1 as to skp the current CU depth and start the predcton from one depth hgher. As for a block belongs to 0, t has a homogenous texture or a smooth gran that follows a certan drecton. In ths way, the block can be easly predcted as one of the 35 predcton modes and results n tny resdual and RD cost. On the other hand, for the block belongs to 1, t owns a vared texture and further splt s requred. The detaled parameters for our tranng s as below. TABLE I TRAINING PARAMETER SETTINGS FOR BLOCK DEPTH 0 AND 1 Parameters Numbers Number of tranng samples 2400 for block depth for block depth 1 Attrbutes per sample (block sze) Number of Classfcaton Max depth 25 Mnmum number of tranng 5 sample for further splt Number of varable randomly selected at node and used to fnd the best splts Number of trees n the forest (=64x64) for block depth 0 2 ( 0 for non-splt, 1 for splt) 64 for block depth 0 32 for block depth 1 The number of tranng samples s set to the extent of a hgh predcton accuracy wthout overfttng. The number of attrbutes per sample, whch means the number of nput feature vectors per sample, s 4096 for mage block sze 64x64 and 1024 for mage block sze 32x32. Then the classfcaton decson s assgned as the last number n a sample, where 0 s for non-splt decson and 1 for splt decson. The maxmum depth of a decson tree s set to 25, whch s obtaned by experments from settng the maxmum depth as 5, 10, 15, The mnmum number of tranng samples for further splt s set to 5. And the number of varables randomly selected at node and used to fnd the best splt s 64 for block sze 64x64, and 32 for block sze 32x32, whch s useful to explore some texture characterstcs. Fnally, we set 10 as the number of trees n the forest. Consderng the optmal parttonng obtaned based on pxel varaton n the ntra-frame predcton has strong dependency wth the scene and the QP value, we traned the classfer sequence by sequence wth four QP values: 22, 27, 32, and 37 as defned n JCT-VC [21]. The ground truth classfcaton s obtaned by encodng the HEVC testng model HM10.0 [22]. Table II and Table III show part of our results about the predcton accuracy on the frst 200 frames s of Class B sequences. From the tables above, we can fnd that APSIPA 1087 APSIPA ASC 2015

4 ths method acheves a hgh predcton accuracy, where most of the results can reach an accuracy of over 90%. The results can guarantee a fast and effcent depth decsons. For the blocks wth wrong classfcaton, there mght be some degradaton n PSNR and ncrease n. But even the sze of CU s wrongly predcted, for example, the ground truth s to stay at depth 0. Our decson stops at depth 2 as shown n Fgure 4. The qualty may not be affected severely because further splt can also acheve an accurate predcton. Wth the low wrong predcton rate, the degradaton of pcture qualty should be low. TABLE II PREDICTION ACCURACY OF CU DEPTH 0 OF CLASS B CU Depth 0 Predcton Average Testng Sequence QP Tranng Basketball BQTerrace Cactus Kmono Park Tme (s) Fg.5 Flowchat of the proposed algorthm confguraton of All-Intra Mode and Man profle (AI-Man). Four QP values (QP 22, QP 27, QP 32, and QP 37) were used to encode all the frames n the test sequences. For each QP of a sequence, we off-lne traned two forests n advance, whch follows our methodology ntroduced n secton The Bjøntegaard delta rate dstorton () performance s measured by Model [23] [24]. The average tme rato () and average tme savng (ATS) are gven by: TABLE III PREDICTION ACCURACY OF CU DEPTH 1 OF CLASS B CU Depth 0 Predcton Average Testng Sequence QP Tranng Basketball BQTerrace Cactus Kmono Park Tme (s) n Enc. Tmeproposed, n Enc. Tme 1 HM10.0, (6) 1 ATS n n Enc. Tme HM10.0, Enc. Tme Enc. Tme 1 HM10.0, proposed, (7) Fg.4 Comparson on CU Sze Decson. (Left: ground truth CU sze encoded by HM10.0. Rght: predctons of our algorthm, all matchng well, except the CU encoded n red.) C. Flowchart of our Proposed Algorthm Fgure 5 sums up our proposed algorthms as llustrated n secton III.A and III.B. When CU encodng starts, the depth range decson wll frstly be found by makng use of neghborng blocks. Then, f the depth level of the current CU s 0 or 1, we use random forests classfers to decde whether the predcton should be bypassed or stayed on the current depth level. For the rest of CU depths (level 2 and 3), conventonal HEVC ntra predcton methods wll be used. IV. EXPERIMENTAL RESULT The proposed fast algorthm s mplemented on the HEVC reference software HM10.0 reference software [22], and was run on a platform wth Intel Core TM CPU 16.0GB RAM sze. Accordng to the common test condtons and reference confguratons recommended by JCT-VC [21], fve groups of testng sequences were carred out to evaluate the codng performance and computatonal complexty under the where n s the total number of sequences, Enc.Tme proposed, s the encodng tme of our proposed algorthm on sequence, and Enc.Tme HM10.0, s the correspondng encodng tme of HM10.0. Table IV summarzes the performance of the overall proposed algorthm. As shown n the tables, the proposed algorthm effcently reduces the computatonal complexty n terms of encodng tme, wth slght ncrease n. Table V gves the comparsons of the proposed algorthm wth some representatve methods n the lterature. For a clear vew of the comparson n graph, we use to reveal the complexty reducton on encodng tme. As shown n Fgure 6, the smaller value n X axs, whch means the lower, the more tme s saved to encode. And the smaller value n Y axs, whch reveals the lower, the less addtonal bts are used to encode. Therefore, the nearer to the left-down corner n the graph, the better s the performance of the algorthm. Our proposed algorthm has a good performance and s comparable to [7] whch s the state-of-the-art result. Furthermore, our proposed algorthm outperforms [8] n both and tme rato, where [8] also has a good result. In addton, [9] s one of very few algorthms that uses machne learnng technque. Although ts s a lttle bt lower than our result, our complexty gans a 20% reducton. Compared wth [25] and [26], our result also has advantages n both and tme savng APSIPA 1088 APSIPA ASC 2015

5 TABLE IV OVERALL RESULTS OF THE PROPOSED ALGORITHM Classfcaton Sequence Num. of frames All Intra Man ATS Traffc Class A (2560x1600) PeopleOnStreet Nebuta SteamLocomotve Kmono Class B (1920x1080) ParkScene Cactus BasketballDrve BQTerrace BasketballDrll Class C (832x480) BQMal PartyScene RaceHorses BasketballPass Class D (416x240) BQSquare BlowngBubbles RaceHorses Class E (1280x720) FourPeople Johnny KrstenAndSara TABLE V PERFORMANCE COMPARISON WITH EXISTING METHOD Class Average Result [7] [8] [9] [25] [26] Proposed all the pxels n the mage block n our random forests scheme and output the decson on whether the process stays n the current depth for predcton or drectly skps the current CU sze and starts the predcton for the hgher depth. Ths s the core strategy of our algorthm whch greatly reduces the computatonal complexty. Expermental result shows that we can acheve the state-of-the-art result, wth 0.80% ncrease and 48.31% tme savng. Fg. 6 Performance comparson wth exstng methods V. CONCLUSION In ths paper, a fast CU depth decson of ntra-frame codng based on a machne learnng technque s proposed. To begn wth, we have used a weghtng sum approach for partton decson. Ths s acheved by usng neghborng CUs as a reference to set the CU sze range for the current CU. Then, we ntroduce random forests to make a predcton on selectng CU sze 64x64 or 32x32. Before encodng, we nput ACKNOWLEDGMENT Ths work s supported by the Center for Sgnal Processng, the Hong Kong Polytechnc Unversty (G-YBAS) and the Research Grant Councl of the Hong Kong SAR Government: PolyU5243/13E(B-Q38S). REFERENCES [1] B. Bross, W.-J. Han, G. J. Sullvan, J.-R. Ohm, and T. Wegand, Hgh effcency vdeo codng (HEVC) text specfcaton draft 6, Jont Collaboratve Team Vdeo Codng (JCT-VC) ITU-T APSIPA 1089 APSIPA ASC 2015

6 VCEG and ISO/IEC MPEG, San Jose, CA, USA, Aprl. 2012, Doc. JCTVC-H1003. [2] T. Wegand, G. J. Sullvan, G. Bjontegaard, and A. Luthra, Overvew of the H.264/AVC vdeo codng standard, IEEE Trans. Crcuts Syst. Vdeo Technol., vol. 13, no. 7, pp , Jul [3] J.-R. Ohm, G. J. Sullvan, H. Schwarz, T. K. Tan, T. Wegand. "Comparson of the codng effcency of vdeo codng standards ncludng hgh effcency vdeo codng (HEVC)." IEEE Trans. Crcuts Syst. Vdeo Technol., vol. 13, no. 12, pp , Jan [4] Mahsa T., Pourazad, C. Doutre, M. Azm, and P. Nasopoulos. "HEVC: the new gold standard for vdeo compresson: how does HEVC compare wth H. 264/AVC?" Consumer Electroncs Magazne, vol. 1, ssue 3, pp , July, [5] D. Marpe, H. Schwarz, and S. Bosse, Vdeo compresson usng nested quadtree structures, leaf mergng, and mproved technques for moton representaton and entropy codng, IEEE Trans. Crcuts Syst. Vdeo Technol., vol. 20, no. 12, pp , Dec [6] J. Lanema, F. Bossen, W. J. Han, J. Mn, and K. Ugur, Intra Codng of the HEVC standard, IEEE Trans. Crcuts Syst. Vdeo Technol., vol. 22, no. 12, pp , Dec [7] S. Cho and M. Km, Fast CU Splttng and Prunng for Suboptmal CU Parttonng n HEVC ntra codng. IEEE Trans. Crcuts Syst. Vdeo Technol., vol. 23, no. 9, pp , Feb [8] L. Shen, Z. Zhang, and Z. Lu. "Effectve CU Sze Decson for HEVC Intracodng." IEEE Trans. on Image Process., vol.23, no.10, pp , July [9] D. Ruz-Coll, V. Adzc, G. Fernandez-Escrbano, H. Kalva, J. L. Martnez, P. Cuenca. Fast Parttonng Algorthm for HEVC Intra Frame Codng Usng Machne Learnng. pp , Proceedngs, IEEE Intern. Confer. on Image Process. (ICIP2014), Pars, France, Oct [10] B. Mn and R. Cheung, A fast CU Sze Decson Algorthm for the HEVC Intra Encoder. IEEE Trans. Crcuts Syst. Vdeo Technol., vol. 25, no. 5, pp , May [11] L.-L. Wang, and W.-C. Su. Novel adaptve algorthm for ntra predcton wth compromsed modes skppng and sgnalng processes n HEVC. IEEE Trans. Crcuts Syst. Vdeo Technol., vol.23, no.13, pp , Sept [12] L.-L. Wang, and W.-C. Su. "H. 264 fast ntra mode selecton algorthm based on drecton dfference measure n the pxel doman." pp , Proceedngs, IEEE Intern. Confer. on Acoustcs, Speech and Sgnal Process. (ICASSP), Tape, Tawan, Aprl [13] Lang Zhao, L Zhang, Swe Ma and Debn Zhao, Fast Mode Decson Algorthm for Intra Predcton n HEVC, pp.1-4, Proceedngs, Vsual Commun. and Image Process. Confer., Tanan, Tawan, Nov [14] J.-H. Mn, S. Lee, I.-K. Km, W.-J. Han, J. Lanema, and K. Ugur, Unfcaton of the Drectonal Intra Predcton Methods n TMuC, Jont Collaboratve Team Vdeo Codng (JCT-VC) ITU-T VCEG and ISO/IEC MPEG, Geneva, Swtzerland, Jul Doc. JCTVC-B100 [15] L. Zhao, L. Zhang, X. Zhao, S. Ma, D. Zhao, W. Gao. "Further Encoder Improvement of ntra mode decson, Jont Collaboratve Team Vdeo Codng (JCT-VC) ITU-T VCEG and ISO/IEC MPEG, Daegu, Korea, Jan. 2011, Doc. JCTVC-D283. [16] Y. Wang, X. Fan, L. Zhao, S. Ma, D. Zhao, W. Gao. A Fast Intra Codng Algorthm for HEVC, pp , Proceedngs, IEEE Intern. Confer. on Image Process. (ICIP2014), Pars, France, Oct [17] L.Breman, Random Forests, Mach. Learn., vol. 45, no.1, pp.5-32, [18] P. Geurts, D. Ernst, and L. Wehenkel, Extremely randomzed trees, Mach. Learn., vol. 63, no.1, pp.3-42, [19] L. Breman, J. Fredman, C. J. Stone, and R.A. Olshen, Classfcaton and Regresson Trees. Boca Raton, FL, USA: CRC Press, [20] A.Crmns, J.Shotton, and E. Konukoglu, Decson Forests: A unfed framework for classfcaton, regresson, densty estmaton, manfold learnng and sem-supervsed learnng, Found. Trends Compt. Graph. Vs., vol. 7, no. 2-3, pp , [21] F. Bossen, Common Test Condtons and Software Reference Confguraton, document JCTVC-L1100, ITU-T/ISO/IEC Jont Collaboratve Team on Vdeo Codng (JCT-VC), 12 th Meetng, Geneve, CH 14-23, Jan [22] Jont Collaboratve Team on Vdeo Codng Reference Software, ver. HM 10.0 [Onlne]. Avalable: svn/svnhevcsoftware/ [23] G. Bjøntegaard, Calculaton of average PSNR dfferences between RD Curves, n Proc. ITU-T SG16/Q.6 VCEG 13th Meet., Austn, TX, USA, Apr. 2001, Doc. VCEG-M33. [24] G. Bjøntegaard, Improvements of the BD-PSNR model, ITU- T SG16/Q.6 VCEG 35th Meet., Berln, Germany, Doc. VCEG-AI11. [25] M.U.K Khan, M. Shafque, J. Henkel, An Adaptve Complexty Reducton Scheme wth Fast Predcton Unt Decson for HEVC Intra Encodng, pp , Proceedngs, IEEE Intern. Confer. on Image Process. (ICIP2013), Melbourne, Australa, Sept [26] J. Km, Y. Choe, and Y.-G. Km, Fast Codng Unt Sze Decson Algorthm for Intra Codng n HEVC, pp , Proceedngs, IEEE Intern. Confer. on Consumer Electroncs (2013ICCE), Berln, Germany, Sept, APSIPA 1090 APSIPA ASC 2015

A Fast Content-Based Multimedia Retrieval Technique Using Compressed Data

A Fast Content-Based Multimedia Retrieval Technique Using Compressed Data A Fast Content-Based Multmeda Retreval Technque Usng Compressed Data Borko Furht and Pornvt Saksobhavvat NSF Multmeda Laboratory Florda Atlantc Unversty, Boca Raton, Florda 3343 ABSTRACT In ths paper,

More information

Fast Intra- and Inter-Prediction Mode Decision in H.264 Advanced Video Coding

Fast Intra- and Inter-Prediction Mode Decision in H.264 Advanced Video Coding Fast Intra- and Inter-Predcton Mode Decson n H.264 Advanced Vdeo Codng Mehd Jafar Islamc Azad Unversty, S and R Branch Department of Communcaton Engneerng P.O.Box 455-775, Tehran, Iran mjafar@mal.uk.ac.r

More information

Fast Intra- and Inter-Prediction Mode Decision in H.264 Advanced Video Coding

Fast Intra- and Inter-Prediction Mode Decision in H.264 Advanced Video Coding 130 IJCSNS Internatonal Journal of Computer Scence and Network Securty, VOL.8 No.5, May 2008 Fast Intra- and Inter-Predcton Mode Decson n H.264 Advanced Vdeo Codng Mehd Jafar and Shohreh Kasae, Islamc

More information

Classifier Selection Based on Data Complexity Measures *

Classifier Selection Based on Data Complexity Measures * Classfer Selecton Based on Data Complexty Measures * Edth Hernández-Reyes, J.A. Carrasco-Ochoa, and J.Fco. Martínez-Trndad Natonal Insttute for Astrophyscs, Optcs and Electroncs, Lus Enrque Erro No.1 Sta.

More information

Learning the Kernel Parameters in Kernel Minimum Distance Classifier

Learning the Kernel Parameters in Kernel Minimum Distance Classifier Learnng the Kernel Parameters n Kernel Mnmum Dstance Classfer Daoqang Zhang 1,, Songcan Chen and Zh-Hua Zhou 1* 1 Natonal Laboratory for Novel Software Technology Nanjng Unversty, Nanjng 193, Chna Department

More information

Improved H.264 Rate Control by Enhanced MAD-Based Frame Complexity Prediction

Improved H.264 Rate Control by Enhanced MAD-Based Frame Complexity Prediction Journal of Vsual Communcaton and Image Representaton (Specal Issue on Emergng H.64/AVC Vdeo Codng Standard), Elsever Scence, May 005 Improved H.64 Rate Control by Enhanced -Based Frame Complexty Predcton

More information

A Binarization Algorithm specialized on Document Images and Photos

A Binarization Algorithm specialized on Document Images and Photos A Bnarzaton Algorthm specalzed on Document mages and Photos Ergna Kavalleratou Dept. of nformaton and Communcaton Systems Engneerng Unversty of the Aegean kavalleratou@aegean.gr Abstract n ths paper, a

More information

Improvement of Spatial Resolution Using BlockMatching Based Motion Estimation and Frame. Integration

Improvement of Spatial Resolution Using BlockMatching Based Motion Estimation and Frame. Integration Improvement of Spatal Resoluton Usng BlockMatchng Based Moton Estmaton and Frame Integraton Danya Suga and Takayuk Hamamoto Graduate School of Engneerng, Tokyo Unversty of Scence, 6-3-1, Nuku, Katsuska-ku,

More information

An Optimal Algorithm for Prufer Codes *

An Optimal Algorithm for Prufer Codes * J. Software Engneerng & Applcatons, 2009, 2: 111-115 do:10.4236/jsea.2009.22016 Publshed Onlne July 2009 (www.scrp.org/journal/jsea) An Optmal Algorthm for Prufer Codes * Xaodong Wang 1, 2, Le Wang 3,

More information

A Fast Visual Tracking Algorithm Based on Circle Pixels Matching

A Fast Visual Tracking Algorithm Based on Circle Pixels Matching A Fast Vsual Trackng Algorthm Based on Crcle Pxels Matchng Zhqang Hou hou_zhq@sohu.com Chongzhao Han czhan@mal.xjtu.edu.cn Ln Zheng Abstract: A fast vsual trackng algorthm based on crcle pxels matchng

More information

Efficient Video Coding with R-D Constrained Quadtree Segmentation

Efficient Video Coding with R-D Constrained Quadtree Segmentation Publshed on Pcture Codng Symposum 1999, March 1999 Effcent Vdeo Codng wth R-D Constraned Quadtree Segmentaton Cha-Wen Ln Computer and Communcaton Research Labs Industral Technology Research Insttute Hsnchu,

More information

Support Vector Machines

Support Vector Machines /9/207 MIST.6060 Busness Intellgence and Data Mnng What are Support Vector Machnes? Support Vector Machnes Support Vector Machnes (SVMs) are supervsed learnng technques that analyze data and recognze patterns.

More information

Term Weighting Classification System Using the Chi-square Statistic for the Classification Subtask at NTCIR-6 Patent Retrieval Task

Term Weighting Classification System Using the Chi-square Statistic for the Classification Subtask at NTCIR-6 Patent Retrieval Task Proceedngs of NTCIR-6 Workshop Meetng, May 15-18, 2007, Tokyo, Japan Term Weghtng Classfcaton System Usng the Ch-square Statstc for the Classfcaton Subtask at NTCIR-6 Patent Retreval Task Kotaro Hashmoto

More information

Parallelism for Nested Loops with Non-uniform and Flow Dependences

Parallelism for Nested Loops with Non-uniform and Flow Dependences Parallelsm for Nested Loops wth Non-unform and Flow Dependences Sam-Jn Jeong Dept. of Informaton & Communcaton Engneerng, Cheonan Unversty, 5, Anseo-dong, Cheonan, Chungnam, 330-80, Korea. seong@cheonan.ac.kr

More information

ALGORITHM FOR H.264/AVC

ALGORITHM FOR H.264/AVC ISSN 1392 124X INFORMATION TECHNOLOGY AND CONTROL, 2010 Vol. 38, No. ISSN 1392 124X INFORMATION TECHNOLOGY AND CONTROL, 2011, Vol.40, No.3 1, 5 7 A NOVEL Novel Varance-Based VARIANCE-BASED Intra-FrameINTRA-FRAME

More information

Fast HEVC Intra Mode Decision Based on Edge Detection and SATD Costs Classification

Fast HEVC Intra Mode Decision Based on Edge Detection and SATD Costs Classification Fast HEVC Intra Mode Decision Based on Edge Detection and SATD Costs Classification Mohammadreza Jamali 1, Stéphane Coulombe 1, François Caron 2 1 École de technologie supérieure, Université du Québec,

More information

Classification Based Mode Decisions for Video over Networks

Classification Based Mode Decisions for Video over Networks Classfcaton Based Mode Decsons for Vdeo over Networks Deepak S. Turaga and Tsuhan Chen Advanced Multmeda Processng Lab Tranng data for Inter-Intra Decson Inter-Intra Decson Regons pdf 6 5 6 5 Energy 4

More information

Skew Angle Estimation and Correction of Hand Written, Textual and Large areas of Non-Textual Document Images: A Novel Approach

Skew Angle Estimation and Correction of Hand Written, Textual and Large areas of Non-Textual Document Images: A Novel Approach Angle Estmaton and Correcton of Hand Wrtten, Textual and Large areas of Non-Textual Document Images: A Novel Approach D.R.Ramesh Babu Pyush M Kumat Mahesh D Dhannawat PES Insttute of Technology Research

More information

Content Based Image Retrieval Using 2-D Discrete Wavelet with Texture Feature with Different Classifiers

Content Based Image Retrieval Using 2-D Discrete Wavelet with Texture Feature with Different Classifiers IOSR Journal of Electroncs and Communcaton Engneerng (IOSR-JECE) e-issn: 78-834,p- ISSN: 78-8735.Volume 9, Issue, Ver. IV (Mar - Apr. 04), PP 0-07 Content Based Image Retreval Usng -D Dscrete Wavelet wth

More information

Feature Reduction and Selection

Feature Reduction and Selection Feature Reducton and Selecton Dr. Shuang LIANG School of Software Engneerng TongJ Unversty Fall, 2012 Today s Topcs Introducton Problems of Dmensonalty Feature Reducton Statstc methods Prncpal Components

More information

EFFICIENT H.264 VIDEO CODING WITH A WORKING MEMORY OF OBJECTS

EFFICIENT H.264 VIDEO CODING WITH A WORKING MEMORY OF OBJECTS EFFICIENT H.264 VIDEO CODING WITH A WORKING MEMORY OF OBJECTS A Thess presented to the Faculty of the Graduate School at the Unversty of Mssour-Columba In Partal Fulfllment of the Requrements for the Degree

More information

Hybrid Non-Blind Color Image Watermarking

Hybrid Non-Blind Color Image Watermarking Hybrd Non-Blnd Color Image Watermarkng Ms C.N.Sujatha 1, Dr. P. Satyanarayana 2 1 Assocate Professor, Dept. of ECE, SNIST, Yamnampet, Ghatkesar Hyderabad-501301, Telangana 2 Professor, Dept. of ECE, AITS,

More information

Combined Rate Control and Mode Decision Optimization for MPEG-2 Transcoding with Spatial Resolution Reduction

Combined Rate Control and Mode Decision Optimization for MPEG-2 Transcoding with Spatial Resolution Reduction MITSUBISHI ELECTRIC RESEARCH LABORATORIES http://www.merl.com Combned Rate Control and Mode Decson Optmzaton for MPEG-2 Transcodng wth Spatal Resoluton Reducton TR2003-7 December 2003 Abstract Ths paper

More information

Classifying Acoustic Transient Signals Using Artificial Intelligence

Classifying Acoustic Transient Signals Using Artificial Intelligence Classfyng Acoustc Transent Sgnals Usng Artfcal Intellgence Steve Sutton, Unversty of North Carolna At Wlmngton (suttons@charter.net) Greg Huff, Unversty of North Carolna At Wlmngton (jgh7476@uncwl.edu)

More information

Problem Definitions and Evaluation Criteria for Computational Expensive Optimization

Problem Definitions and Evaluation Criteria for Computational Expensive Optimization Problem efntons and Evaluaton Crtera for Computatonal Expensve Optmzaton B. Lu 1, Q. Chen and Q. Zhang 3, J. J. Lang 4, P. N. Suganthan, B. Y. Qu 6 1 epartment of Computng, Glyndwr Unversty, UK Faclty

More information

Support Vector Machines

Support Vector Machines Support Vector Machnes Decson surface s a hyperplane (lne n 2D) n feature space (smlar to the Perceptron) Arguably, the most mportant recent dscovery n machne learnng In a nutshell: map the data to a predetermned

More information

Helsinki University Of Technology, Systems Analysis Laboratory Mat Independent research projects in applied mathematics (3 cr)

Helsinki University Of Technology, Systems Analysis Laboratory Mat Independent research projects in applied mathematics (3 cr) Helsnk Unversty Of Technology, Systems Analyss Laboratory Mat-2.08 Independent research projects n appled mathematcs (3 cr) "! #$&% Antt Laukkanen 506 R ajlaukka@cc.hut.f 2 Introducton...3 2 Multattrbute

More information

Shape-adaptive DCT and Its Application in Region-based Image Coding

Shape-adaptive DCT and Its Application in Region-based Image Coding Internatonal Journal of Sgnal Processng, Image Processng and Pattern Recognton, pp.99-108 http://dx.do.org/10.14257/sp.2014.7.1.10 Shape-adaptve DCT and Its Applcaton n Regon-based Image Codng Yamn Zheng,

More information

Collaboratively Regularized Nearest Points for Set Based Recognition

Collaboratively Regularized Nearest Points for Set Based Recognition Academc Center for Computng and Meda Studes, Kyoto Unversty Collaboratvely Regularzed Nearest Ponts for Set Based Recognton Yang Wu, Mchhko Mnoh, Masayuk Mukunok Kyoto Unversty 9/1/013 BMVC 013 @ Brstol,

More information

MULTISPECTRAL IMAGES CLASSIFICATION BASED ON KLT AND ATR AUTOMATIC TARGET RECOGNITION

MULTISPECTRAL IMAGES CLASSIFICATION BASED ON KLT AND ATR AUTOMATIC TARGET RECOGNITION MULTISPECTRAL IMAGES CLASSIFICATION BASED ON KLT AND ATR AUTOMATIC TARGET RECOGNITION Paulo Quntlano 1 & Antono Santa-Rosa 1 Federal Polce Department, Brasla, Brazl. E-mals: quntlano.pqs@dpf.gov.br and

More information

BOOSTING CLASSIFICATION ACCURACY WITH SAMPLES CHOSEN FROM A VALIDATION SET

BOOSTING CLASSIFICATION ACCURACY WITH SAMPLES CHOSEN FROM A VALIDATION SET 1 BOOSTING CLASSIFICATION ACCURACY WITH SAMPLES CHOSEN FROM A VALIDATION SET TZU-CHENG CHUANG School of Electrcal and Computer Engneerng, Purdue Unversty, West Lafayette, Indana 47907 SAUL B. GELFAND School

More information

Face Recognition University at Buffalo CSE666 Lecture Slides Resources:

Face Recognition University at Buffalo CSE666 Lecture Slides Resources: Face Recognton Unversty at Buffalo CSE666 Lecture Sldes Resources: http://www.face-rec.org/algorthms/ Overvew of face recognton algorthms Correlaton - Pxel based correspondence between two face mages Structural

More information

FEATURE EXTRACTION. Dr. K.Vijayarekha. Associate Dean School of Electrical and Electronics Engineering SASTRA University, Thanjavur

FEATURE EXTRACTION. Dr. K.Vijayarekha. Associate Dean School of Electrical and Electronics Engineering SASTRA University, Thanjavur FEATURE EXTRACTION Dr. K.Vjayarekha Assocate Dean School of Electrcal and Electroncs Engneerng SASTRA Unversty, Thanjavur613 41 Jont Intatve of IITs and IISc Funded by MHRD Page 1 of 8 Table of Contents

More information

A Background Subtraction for a Vision-based User Interface *

A Background Subtraction for a Vision-based User Interface * A Background Subtracton for a Vson-based User Interface * Dongpyo Hong and Woontack Woo KJIST U-VR Lab. {dhon wwoo}@kjst.ac.kr Abstract In ths paper, we propose a robust and effcent background subtracton

More information

EYE CENTER LOCALIZATION ON A FACIAL IMAGE BASED ON MULTI-BLOCK LOCAL BINARY PATTERNS

EYE CENTER LOCALIZATION ON A FACIAL IMAGE BASED ON MULTI-BLOCK LOCAL BINARY PATTERNS P.G. Demdov Yaroslavl State Unversty Anatoly Ntn, Vladmr Khryashchev, Olga Stepanova, Igor Kostern EYE CENTER LOCALIZATION ON A FACIAL IMAGE BASED ON MULTI-BLOCK LOCAL BINARY PATTERNS Yaroslavl, 2015 Eye

More information

Fuzzy Modeling of the Complexity vs. Accuracy Trade-off in a Sequential Two-Stage Multi-Classifier System

Fuzzy Modeling of the Complexity vs. Accuracy Trade-off in a Sequential Two-Stage Multi-Classifier System Fuzzy Modelng of the Complexty vs. Accuracy Trade-off n a Sequental Two-Stage Mult-Classfer System MARK LAST 1 Department of Informaton Systems Engneerng Ben-Guron Unversty of the Negev Beer-Sheva 84105

More information

Determining the Optimal Bandwidth Based on Multi-criterion Fusion

Determining the Optimal Bandwidth Based on Multi-criterion Fusion Proceedngs of 01 4th Internatonal Conference on Machne Learnng and Computng IPCSIT vol. 5 (01) (01) IACSIT Press, Sngapore Determnng the Optmal Bandwdth Based on Mult-crteron Fuson Ha-L Lang 1+, Xan-Mn

More information

Edge Detection in Noisy Images Using the Support Vector Machines

Edge Detection in Noisy Images Using the Support Vector Machines Edge Detecton n Nosy Images Usng the Support Vector Machnes Hlaro Gómez-Moreno, Saturnno Maldonado-Bascón, Francsco López-Ferreras Sgnal Theory and Communcatons Department. Unversty of Alcalá Crta. Madrd-Barcelona

More information

Outline. Discriminative classifiers for image recognition. Where in the World? A nearest neighbor recognition example 4/14/2011. CS 376 Lecture 22 1

Outline. Discriminative classifiers for image recognition. Where in the World? A nearest neighbor recognition example 4/14/2011. CS 376 Lecture 22 1 4/14/011 Outlne Dscrmnatve classfers for mage recognton Wednesday, Aprl 13 Krsten Grauman UT-Austn Last tme: wndow-based generc obect detecton basc ppelne face detecton wth boostng as case study Today:

More information

A Modified Median Filter for the Removal of Impulse Noise Based on the Support Vector Machines

A Modified Median Filter for the Removal of Impulse Noise Based on the Support Vector Machines A Modfed Medan Flter for the Removal of Impulse Nose Based on the Support Vector Machnes H. GOMEZ-MORENO, S. MALDONADO-BASCON, F. LOPEZ-FERRERAS, M. UTRILLA- MANSO AND P. GIL-JIMENEZ Departamento de Teoría

More information

Incremental Learning with Support Vector Machines and Fuzzy Set Theory

Incremental Learning with Support Vector Machines and Fuzzy Set Theory The 25th Workshop on Combnatoral Mathematcs and Computaton Theory Incremental Learnng wth Support Vector Machnes and Fuzzy Set Theory Yu-Mng Chuang 1 and Cha-Hwa Ln 2* 1 Department of Computer Scence and

More information

Enhanced AMBTC for Image Compression using Block Classification and Interpolation

Enhanced AMBTC for Image Compression using Block Classification and Interpolation Internatonal Journal of Computer Applcatons (0975 8887) Volume 5 No.0, August 0 Enhanced AMBTC for Image Compresson usng Block Classfcaton and Interpolaton S. Vmala Dept. of Comp. Scence Mother Teresa

More information

A PATTERN RECOGNITION APPROACH TO IMAGE SEGMENTATION

A PATTERN RECOGNITION APPROACH TO IMAGE SEGMENTATION 1 THE PUBLISHING HOUSE PROCEEDINGS OF THE ROMANIAN ACADEMY, Seres A, OF THE ROMANIAN ACADEMY Volume 4, Number 2/2003, pp.000-000 A PATTERN RECOGNITION APPROACH TO IMAGE SEGMENTATION Tudor BARBU Insttute

More information

Shape Representation Robust to the Sketching Order Using Distance Map and Direction Histogram

Shape Representation Robust to the Sketching Order Using Distance Map and Direction Histogram Shape Representaton Robust to the Sketchng Order Usng Dstance Map and Drecton Hstogram Department of Computer Scence Yonse Unversty Kwon Yun CONTENTS Revew Topc Proposed Method System Overvew Sketch Normalzaton

More information

Detection of Double AVC/HEVC Encoding

Detection of Double AVC/HEVC Encoding 26 24th European Sgnal Processng Conference (EUSIPCO) Detecton of Double AVC/HEVC Encodng Andrea Costanzo, Mauro Barn, Fellow, IEEE Dept. of Informaton Engneerng and Mathematcs, Unversty of Sena, Italy

More information

The Research of Support Vector Machine in Agricultural Data Classification

The Research of Support Vector Machine in Agricultural Data Classification The Research of Support Vector Machne n Agrcultural Data Classfcaton Le Sh, Qguo Duan, Xnmng Ma, Me Weng College of Informaton and Management Scence, HeNan Agrcultural Unversty, Zhengzhou 45000 Chna Zhengzhou

More information

A Novel Adaptive Descriptor Algorithm for Ternary Pattern Textures

A Novel Adaptive Descriptor Algorithm for Ternary Pattern Textures A Novel Adaptve Descrptor Algorthm for Ternary Pattern Textures Fahuan Hu 1,2, Guopng Lu 1 *, Zengwen Dong 1 1.School of Mechancal & Electrcal Engneerng, Nanchang Unversty, Nanchang, 330031, Chna; 2. School

More information

An Image Fusion Approach Based on Segmentation Region

An Image Fusion Approach Based on Segmentation Region Rong Wang, L-Qun Gao, Shu Yang, Yu-Hua Cha, and Yan-Chun Lu An Image Fuson Approach Based On Segmentaton Regon An Image Fuson Approach Based on Segmentaton Regon Rong Wang, L-Qun Gao, Shu Yang 3, Yu-Hua

More information

Journal of Chemical and Pharmaceutical Research, 2014, 6(6): Research Article. A selective ensemble classification method on microarray data

Journal of Chemical and Pharmaceutical Research, 2014, 6(6): Research Article. A selective ensemble classification method on microarray data Avalable onlne www.ocpr.com Journal of Chemcal and Pharmaceutcal Research, 2014, 6(6):2860-2866 Research Artcle ISSN : 0975-7384 CODEN(USA) : JCPRC5 A selectve ensemble classfcaton method on mcroarray

More information

Three supervised learning methods on pen digits character recognition dataset

Three supervised learning methods on pen digits character recognition dataset Three supervsed learnng methods on pen dgts character recognton dataset Chrs Flezach Department of Computer Scence and Engneerng Unversty of Calforna, San Dego San Dego, CA 92093 cflezac@cs.ucsd.edu Satoru

More information

Cluster Analysis of Electrical Behavior

Cluster Analysis of Electrical Behavior Journal of Computer and Communcatons, 205, 3, 88-93 Publshed Onlne May 205 n ScRes. http://www.scrp.org/ournal/cc http://dx.do.org/0.4236/cc.205.350 Cluster Analyss of Electrcal Behavor Ln Lu Ln Lu, School

More information

Lecture 5: Multilayer Perceptrons

Lecture 5: Multilayer Perceptrons Lecture 5: Multlayer Perceptrons Roger Grosse 1 Introducton So far, we ve only talked about lnear models: lnear regresson and lnear bnary classfers. We noted that there are functons that can t be represented

More information

A Statistical Model Selection Strategy Applied to Neural Networks

A Statistical Model Selection Strategy Applied to Neural Networks A Statstcal Model Selecton Strategy Appled to Neural Networks Joaquín Pzarro Elsa Guerrero Pedro L. Galndo joaqun.pzarro@uca.es elsa.guerrero@uca.es pedro.galndo@uca.es Dpto Lenguajes y Sstemas Informátcos

More information

Fast Feature Value Searching for Face Detection

Fast Feature Value Searching for Face Detection Vol., No. 2 Computer and Informaton Scence Fast Feature Value Searchng for Face Detecton Yunyang Yan Department of Computer Engneerng Huayn Insttute of Technology Hua an 22300, Chna E-mal: areyyyke@63.com

More information

Lobachevsky State University of Nizhni Novgorod. Polyhedron. Quick Start Guide

Lobachevsky State University of Nizhni Novgorod. Polyhedron. Quick Start Guide Lobachevsky State Unversty of Nzhn Novgorod Polyhedron Quck Start Gude Nzhn Novgorod 2016 Contents Specfcaton of Polyhedron software... 3 Theoretcal background... 4 1. Interface of Polyhedron... 6 1.1.

More information

CHAPTER 3 SEQUENTIAL MINIMAL OPTIMIZATION TRAINED SUPPORT VECTOR CLASSIFIER FOR CANCER PREDICTION

CHAPTER 3 SEQUENTIAL MINIMAL OPTIMIZATION TRAINED SUPPORT VECTOR CLASSIFIER FOR CANCER PREDICTION 48 CHAPTER 3 SEQUENTIAL MINIMAL OPTIMIZATION TRAINED SUPPORT VECTOR CLASSIFIER FOR CANCER PREDICTION 3.1 INTRODUCTION The raw mcroarray data s bascally an mage wth dfferent colors ndcatng hybrdzaton (Xue

More information

A DCVS Reconstruction Algorithm for Mine Video Monitoring Image Based on Block Classification

A DCVS Reconstruction Algorithm for Mine Video Monitoring Image Based on Block Classification 1 3 4 5 6 7 8 9 10 11 1 13 14 15 16 17 18 19 0 1 Artcle A DCVS Reconstructon Algorthm for Mne Vdeo Montorng Image Based on Block Classfcaton Xaohu Zhao 1,, Xueru Shen 1,, *, Kuan Wang 1, and Wanme L 1,

More information

Virtual Machine Migration based on Trust Measurement of Computer Node

Virtual Machine Migration based on Trust Measurement of Computer Node Appled Mechancs and Materals Onlne: 2014-04-04 ISSN: 1662-7482, Vols. 536-537, pp 678-682 do:10.4028/www.scentfc.net/amm.536-537.678 2014 Trans Tech Publcatons, Swtzerland Vrtual Machne Mgraton based on

More information

On Modeling Variations For Face Authentication

On Modeling Variations For Face Authentication On Modelng Varatons For Face Authentcaton Xaomng Lu Tsuhan Chen B.V.K. Vjaya Kumar Department of Electrcal and Computer Engneerng, Carnege Mellon Unversty Abstract In ths paper, we present a scheme for

More information

User Authentication Based On Behavioral Mouse Dynamics Biometrics

User Authentication Based On Behavioral Mouse Dynamics Biometrics User Authentcaton Based On Behavoral Mouse Dynamcs Bometrcs Chee-Hyung Yoon Danel Donghyun Km Department of Computer Scence Department of Computer Scence Stanford Unversty Stanford Unversty Stanford, CA

More information

An Image Compression Algorithm based on Wavelet Transform and LZW

An Image Compression Algorithm based on Wavelet Transform and LZW An Image Compresson Algorthm based on Wavelet Transform and LZW Png Luo a, Janyong Yu b School of Chongqng Unversty of Posts and Telecommuncatons, Chongqng, 400065, Chna Abstract a cylpng@63.com, b y27769864@sna.cn

More information

Load Balancing for Hex-Cell Interconnection Network

Load Balancing for Hex-Cell Interconnection Network Int. J. Communcatons, Network and System Scences,,, - Publshed Onlne Aprl n ScRes. http://www.scrp.org/journal/jcns http://dx.do.org/./jcns.. Load Balancng for Hex-Cell Interconnecton Network Saher Manaseer,

More information

Available online at Available online at Advanced in Control Engineering and Information Science

Available online at   Available online at   Advanced in Control Engineering and Information Science Avalable onlne at wwwscencedrectcom Avalable onlne at wwwscencedrectcom Proceda Proceda Engneerng Engneerng 00 (2011) 15000 000 (2011) 1642 1646 Proceda Engneerng wwwelsevercom/locate/proceda Advanced

More information

Course Introduction. Algorithm 8/31/2017. COSC 320 Advanced Data Structures and Algorithms. COSC 320 Advanced Data Structures and Algorithms

Course Introduction. Algorithm 8/31/2017. COSC 320 Advanced Data Structures and Algorithms. COSC 320 Advanced Data Structures and Algorithms Course Introducton Course Topcs Exams, abs, Proects A quc loo at a few algorthms 1 Advanced Data Structures and Algorthms Descrpton: We are gong to dscuss algorthm complexty analyss, algorthm desgn technques

More information

Conditional Speculative Decimal Addition*

Conditional Speculative Decimal Addition* Condtonal Speculatve Decmal Addton Alvaro Vazquez and Elsardo Antelo Dep. of Electronc and Computer Engneerng Unv. of Santago de Compostela, Span Ths work was supported n part by Xunta de Galca under grant

More information

Novel Fuzzy logic Based Edge Detection Technique

Novel Fuzzy logic Based Edge Detection Technique Novel Fuzzy logc Based Edge Detecton Technque Aborsade, D.O Department of Electroncs Engneerng, adoke Akntola Unversty of Tech., Ogbomoso. Oyo-state. doaborsade@yahoo.com Abstract Ths paper s based on

More information

Under-Sampling Approaches for Improving Prediction of the Minority Class in an Imbalanced Dataset

Under-Sampling Approaches for Improving Prediction of the Minority Class in an Imbalanced Dataset Under-Samplng Approaches for Improvng Predcton of the Mnorty Class n an Imbalanced Dataset Show-Jane Yen and Yue-Sh Lee Department of Computer Scence and Informaton Engneerng, Mng Chuan Unversty 5 The-Mng

More information

An efficient method to build panoramic image mosaics

An efficient method to build panoramic image mosaics An effcent method to buld panoramc mage mosacs Pattern Recognton Letters vol. 4 003 Dae-Hyun Km Yong-In Yoon Jong-Soo Cho School of Electrcal Engneerng and Computer Scence Kyungpook Natonal Unv. Abstract

More information

A Lazy Ensemble Learning Method to Classification

A Lazy Ensemble Learning Method to Classification IJCSI Internatonal Journal of Computer Scence Issues, Vol. 7, Issue 5, September 2010 ISSN (Onlne): 1694-0814 344 A Lazy Ensemble Learnng Method to Classfcaton Haleh Homayoun 1, Sattar Hashem 2 and Al

More information

Detection of an Object by using Principal Component Analysis

Detection of an Object by using Principal Component Analysis Detecton of an Object by usng Prncpal Component Analyss 1. G. Nagaven, 2. Dr. T. Sreenvasulu Reddy 1. M.Tech, Department of EEE, SVUCE, Trupath, Inda. 2. Assoc. Professor, Department of ECE, SVUCE, Trupath,

More information

CLASSIFICATION OF ULTRASONIC SIGNALS

CLASSIFICATION OF ULTRASONIC SIGNALS The 8 th Internatonal Conference of the Slovenan Socety for Non-Destructve Testng»Applcaton of Contemporary Non-Destructve Testng n Engneerng«September -3, 5, Portorož, Slovena, pp. 7-33 CLASSIFICATION

More information

WITH the rapid development of visual communication

WITH the rapid development of visual communication IEEE TRASACTIOS O IMAGE PROCESSIG, VOL. 25, O. 7, JULY 2016 2943 Low-Delay Rate Control for Consstent Qualty Usng Dstorton-Based Lagrange Multpler Maohu Wang, Member, IEEE, Kngggan,Fellow, IEEE, and Honglang

More information

BLOCK-ADAPTIVE INTERPOLATION FILTER FOR SUB-PIXEL MOTION COMPENSATION

BLOCK-ADAPTIVE INTERPOLATION FILTER FOR SUB-PIXEL MOTION COMPENSATION 19th European Sgnal rocessng Conference (EUSICO 2011) Barcelona, Span, August 29 - September 2, 2011 BOCK-ADATIVE INTEROATION FITER FOR SUB-IXE MOTION COMENSATION Jaehyun Cho, Dong-Bo ee, Shn Cheol Jeong,

More information

The Greedy Method. Outline and Reading. Change Money Problem. Greedy Algorithms. Applications of the Greedy Strategy. The Greedy Method Technique

The Greedy Method. Outline and Reading. Change Money Problem. Greedy Algorithms. Applications of the Greedy Strategy. The Greedy Method Technique //00 :0 AM Outlne and Readng The Greedy Method The Greedy Method Technque (secton.) Fractonal Knapsack Problem (secton..) Task Schedulng (secton..) Mnmum Spannng Trees (secton.) Change Money Problem Greedy

More information

Video Content Representation using Optimal Extraction of Frames and Scenes

Video Content Representation using Optimal Extraction of Frames and Scenes Vdeo Content Representaton usng Optmal Etracton of rames and Scenes Nkolaos D. Doulam Anastasos D. Doulam Yanns S. Avrths and Stefanos D. ollas Natonal Techncal Unversty of Athens Department of Electrcal

More information

An Approach to Selective Intra Coding and Early Inter Skip Prediction in H.264/AVC Standard

An Approach to Selective Intra Coding and Early Inter Skip Prediction in H.264/AVC Standard FACTA UNIVERSITATIS (NIŠ) SER.: ELEC. ENERG. vol. 21, no. 1, Aprl 2008, 107-119 An Approach to Selectve Intra Codng and Early Inter Skp Predcton n H.264/AVC Standard Zoran Mlčevć and Zoran Bojkovć Abstract:

More information

Machine Learning: Algorithms and Applications

Machine Learning: Algorithms and Applications 14/05/1 Machne Learnng: Algorthms and Applcatons Florano Zn Free Unversty of Bozen-Bolzano Faculty of Computer Scence Academc Year 011-01 Lecture 10: 14 May 01 Unsupervsed Learnng cont Sldes courtesy of

More information

NUMERICAL SOLVING OPTIMAL CONTROL PROBLEMS BY THE METHOD OF VARIATIONS

NUMERICAL SOLVING OPTIMAL CONTROL PROBLEMS BY THE METHOD OF VARIATIONS ARPN Journal of Engneerng and Appled Scences 006-017 Asan Research Publshng Network (ARPN). All rghts reserved. NUMERICAL SOLVING OPTIMAL CONTROL PROBLEMS BY THE METHOD OF VARIATIONS Igor Grgoryev, Svetlana

More information

The Study of Remote Sensing Image Classification Based on Support Vector Machine

The Study of Remote Sensing Image Classification Based on Support Vector Machine Sensors & Transducers 03 by IFSA http://www.sensorsportal.com The Study of Remote Sensng Image Classfcaton Based on Support Vector Machne, ZHANG Jan-Hua Key Research Insttute of Yellow Rver Cvlzaton and

More information

PERFORMANCE EVALUATION FOR SCENE MATCHING ALGORITHMS BY SVM

PERFORMANCE EVALUATION FOR SCENE MATCHING ALGORITHMS BY SVM PERFORMACE EVALUAIO FOR SCEE MACHIG ALGORIHMS BY SVM Zhaohu Yang a, b, *, Yngyng Chen a, Shaomng Zhang a a he Research Center of Remote Sensng and Geomatc, ongj Unversty, Shangha 200092, Chna - yzhac@63.com

More information

Discriminative Dictionary Learning with Pairwise Constraints

Discriminative Dictionary Learning with Pairwise Constraints Dscrmnatve Dctonary Learnng wth Parwse Constrants Humn Guo Zhuoln Jang LARRY S. DAVIS UNIVERSITY OF MARYLAND Nov. 6 th, Outlne Introducton/motvaton Dctonary Learnng Dscrmnatve Dctonary Learnng wth Parwse

More information

Recommended Items Rating Prediction based on RBF Neural Network Optimized by PSO Algorithm

Recommended Items Rating Prediction based on RBF Neural Network Optimized by PSO Algorithm Recommended Items Ratng Predcton based on RBF Neural Network Optmzed by PSO Algorthm Chengfang Tan, Cayn Wang, Yuln L and Xx Q Abstract In order to mtgate the data sparsty and cold-start problems of recommendaton

More information

SCALABLE AND VISUALIZATION-ORIENTED CLUSTERING FOR EXPLORATORY SPATIAL ANALYSIS

SCALABLE AND VISUALIZATION-ORIENTED CLUSTERING FOR EXPLORATORY SPATIAL ANALYSIS SCALABLE AND VISUALIZATION-ORIENTED CLUSTERING FOR EXPLORATORY SPATIAL ANALYSIS J.H.Guan, F.B.Zhu, F.L.Ban a School of Computer, Spatal Informaton & Dgtal Engneerng Center, Wuhan Unversty, Wuhan, 430079,

More information

Object-Based Techniques for Image Retrieval

Object-Based Techniques for Image Retrieval 54 Zhang, Gao, & Luo Chapter VII Object-Based Technques for Image Retreval Y. J. Zhang, Tsnghua Unversty, Chna Y. Y. Gao, Tsnghua Unversty, Chna Y. Luo, Tsnghua Unversty, Chna ABSTRACT To overcome the

More information

An Improved Image Segmentation Algorithm Based on the Otsu Method

An Improved Image Segmentation Algorithm Based on the Otsu Method 3th ACIS Internatonal Conference on Software Engneerng, Artfcal Intellgence, Networkng arallel/dstrbuted Computng An Improved Image Segmentaton Algorthm Based on the Otsu Method Mengxng Huang, enjao Yu,

More information

6.854 Advanced Algorithms Petar Maymounkov Problem Set 11 (November 23, 2005) With: Benjamin Rossman, Oren Weimann, and Pouya Kheradpour

6.854 Advanced Algorithms Petar Maymounkov Problem Set 11 (November 23, 2005) With: Benjamin Rossman, Oren Weimann, and Pouya Kheradpour 6.854 Advanced Algorthms Petar Maymounkov Problem Set 11 (November 23, 2005) Wth: Benjamn Rossman, Oren Wemann, and Pouya Kheradpour Problem 1. We reduce vertex cover to MAX-SAT wth weghts, such that the

More information

Joint Example-based Depth Map Super-Resolution

Joint Example-based Depth Map Super-Resolution Jont Example-based Depth Map Super-Resoluton Yanje L 1, Tanfan Xue,3, Lfeng Sun 1, Janzhuang Lu,3,4 1 Informaton Scence and Technology Department, Tsnghua Unversty, Bejng, Chna Department of Informaton

More information

BIN XIA et al: AN IMPROVED K-MEANS ALGORITHM BASED ON CLOUD PLATFORM FOR DATA MINING

BIN XIA et al: AN IMPROVED K-MEANS ALGORITHM BASED ON CLOUD PLATFORM FOR DATA MINING An Improved K-means Algorthm based on Cloud Platform for Data Mnng Bn Xa *, Yan Lu 2. School of nformaton and management scence, Henan Agrcultural Unversty, Zhengzhou, Henan 450002, P.R. Chna 2. College

More information

Associative Based Classification Algorithm For Diabetes Disease Prediction

Associative Based Classification Algorithm For Diabetes Disease Prediction Internatonal Journal of Engneerng Trends and Technology (IJETT) Volume-41 Number-3 - November 016 Assocatve Based Classfcaton Algorthm For Dabetes Dsease Predcton 1 N. Gnana Deepka, Y.surekha, 3 G.Laltha

More information

A Unified Framework for Semantics and Feature Based Relevance Feedback in Image Retrieval Systems

A Unified Framework for Semantics and Feature Based Relevance Feedback in Image Retrieval Systems A Unfed Framework for Semantcs and Feature Based Relevance Feedback n Image Retreval Systems Ye Lu *, Chunhu Hu 2, Xngquan Zhu 3*, HongJang Zhang 2, Qang Yang * School of Computng Scence Smon Fraser Unversty

More information

SHAPE RECOGNITION METHOD BASED ON THE k-nearest NEIGHBOR RULE

SHAPE RECOGNITION METHOD BASED ON THE k-nearest NEIGHBOR RULE SHAPE RECOGNITION METHOD BASED ON THE k-nearest NEIGHBOR RULE Dorna Purcaru Faculty of Automaton, Computers and Electroncs Unersty of Craoa 13 Al. I. Cuza Street, Craoa RO-1100 ROMANIA E-mal: dpurcaru@electroncs.uc.ro

More information

TECHNIQUE OF FORMATION HOMOGENEOUS SAMPLE SAME OBJECTS. Muradaliyev A.Z.

TECHNIQUE OF FORMATION HOMOGENEOUS SAMPLE SAME OBJECTS. Muradaliyev A.Z. TECHNIQUE OF FORMATION HOMOGENEOUS SAMPLE SAME OBJECTS Muradalyev AZ Azerbajan Scentfc-Research and Desgn-Prospectng Insttute of Energetc AZ1012, Ave HZardab-94 E-mal:aydn_murad@yahoocom Importance of

More information

S1 Note. Basis functions.

S1 Note. Basis functions. S1 Note. Bass functons. Contents Types of bass functons...1 The Fourer bass...2 B-splne bass...3 Power and type I error rates wth dfferent numbers of bass functons...4 Table S1. Smulaton results of type

More information

Local Quaternary Patterns and Feature Local Quaternary Patterns

Local Quaternary Patterns and Feature Local Quaternary Patterns Local Quaternary Patterns and Feature Local Quaternary Patterns Jayu Gu and Chengjun Lu The Department of Computer Scence, New Jersey Insttute of Technology, Newark, NJ 0102, USA Abstract - Ths paper presents

More information

Fuzzy Filtering Algorithms for Image Processing: Performance Evaluation of Various Approaches

Fuzzy Filtering Algorithms for Image Processing: Performance Evaluation of Various Approaches Proceedngs of the Internatonal Conference on Cognton and Recognton Fuzzy Flterng Algorthms for Image Processng: Performance Evaluaton of Varous Approaches Rajoo Pandey and Umesh Ghanekar Department of

More information

A fast algorithm for color image segmentation

A fast algorithm for color image segmentation Unersty of Wollongong Research Onlne Faculty of Informatcs - Papers (Arche) Faculty of Engneerng and Informaton Scences 006 A fast algorthm for color mage segmentaton L. Dong Unersty of Wollongong, lju@uow.edu.au

More information

Outline. Type of Machine Learning. Examples of Application. Unsupervised Learning

Outline. Type of Machine Learning. Examples of Application. Unsupervised Learning Outlne Artfcal Intellgence and ts applcatons Lecture 8 Unsupervsed Learnng Professor Danel Yeung danyeung@eee.org Dr. Patrck Chan patrckchan@eee.org South Chna Unversty of Technology, Chna Introducton

More information

Smoothing Spline ANOVA for variable screening

Smoothing Spline ANOVA for variable screening Smoothng Splne ANOVA for varable screenng a useful tool for metamodels tranng and mult-objectve optmzaton L. Rcco, E. Rgon, A. Turco Outlne RSM Introducton Possble couplng Test case MOO MOO wth Game Theory

More information

Efficient Segmentation and Classification of Remote Sensing Image Using Local Self Similarity

Efficient Segmentation and Classification of Remote Sensing Image Using Local Self Similarity ISSN(Onlne): 2320-9801 ISSN (Prnt): 2320-9798 Internatonal Journal of Innovatve Research n Computer and Communcaton Engneerng (An ISO 3297: 2007 Certfed Organzaton) Vol.2, Specal Issue 1, March 2014 Proceedngs

More information

FRACTAL COMPRESSION TECHNIQUE FOR COLOR IMAGES USING VARIABLE BLOCK

FRACTAL COMPRESSION TECHNIQUE FOR COLOR IMAGES USING VARIABLE BLOCK ISSN: 0976-910 (ONLINE) ICTACT JOURNAL ON IMAGE AND VIDEO PROCESSING, NOVEMBER 017, VOLUME: 08, ISSUE: 0 DOI: 10.1917/jvp.017.030 FRACTAL COMPRESSION TECHNIQUE FOR COLOR IMAGES USING VARIABLE BLOCK Nsar

More information