High Precision Gait Recognition Using a Large-Scale PC Cluster

Size: px
Start display at page:

Download "High Precision Gait Recognition Using a Large-Scale PC Cluster"

Transcription

1 High Precision Gait Recognition Using a Large-Scale PC Cluster Yuuki Horita 1 Satoshi Ito 1 Kenji Kaneda 1 Takuya Nanri 1 Yasuyuki Shimohata 1 Kenjiro Taura 1 Mihoko Otake 2 Tomomasa Sato 1 Nobuyuki Otsu Graduate School of Information Science and Technology, University of Tokyo Hongo, Bunkyo-ku, Tokyo , Japan. 2 Division of Project Coordination, University of Tokyo Kashiwa-no-ha, Kashiwa, Chiba , Japan. 3 National Institute of Advanced Industrial Science and Technology Umezono, Tsukuba, Ibaraki , Japan. Abstract. A high-precision gait recognition method is vital for good video surveillance systems. Gait recognition methods proposed so far include a method with cubic higher-order local auto-correlation (CHLAC) features. However, the recognition rate of this method is lowered if the size of objects to be recognized varies dynamically. Although one possible solution to the problem is to increase the number of CHLAC features, it is not a trivial matter to implement the method because of its large amount of computation time. To overcome this problem, we propose a method that uses a larger number of CHLAC features than a traditional CHLAC method by employing parallel computing techniques. We evaluated our method by applying it to a data set containing objects whose scale changes dynamically. We ran the parallel version of the program on a 192-node cluster. The experimental results show that our method achieved a higher recognition rate than that of the traditional CHLAC method. The results also show that the parallel version of the program achieved a better speedup than the sequential version. Keywords: Gait Recognition, Cubic Higher-order Local Auto-Correlation, Feature Extraction, Cluster Computing, Parallel and Distributed Algorithms 1 Introduction With a recent increase in crimes and terrorist attacks, video-surveillance systems (e.g., video monitoring and automatic recognition techniques) are becoming more and more essential to provide a safe society. For example, the video-surveillance systems are useful in cases where criminals and terrorists usually hide their faces. While it is difficult to recognize them with their face images obtained by cameras, the video-surveillance systems can achieve a high recognition rate since the systems identify individuals using information about their movements, which are difficult for people to hide. Since recognition errors made by video-surveillance systems cause critical damage to a safe society, features enabling gait recognition methods to achieve a high recognition rate have been proposed so far. Many methods based on silhouettes of a human figure were proposed [6, 12, 15]. However, these methods need segmentation, so it is easy to be affected by noise. The cubic higherorder local auto-correlation (CHLAC) feature [7], which has been proposed by Kobayashi and Otsu, is one the features that have shown a high recognition rate. For example, a gait recognition method with the CHLAC features achieved the highest recognition rate when it was applied to data sets of HumanID Gait Challenge Problem [12], which was designed for international competitions [8]. However, the recognition rate of the existing method {horita, tau}@logos.ic.i.u-tokyo.ac.jp {itou, nanri, simohata}@isi.imi.i.u-tokyo.ac.jp kaneda@yl.is.s.u-tokyo.ac.jp otake@cb.k.u-tokyo.ac.jp tomo@ics.t.u-tokyo.ac.jp otsu.n@aist.go.jp with CHLAC features goes down when the given data sets are more complex. For example, the method faces difficulty in recognizing data sets that contain various kinds of movements and scale variation of target objects. Although one possible solution to achieve a higher recognition rate is to increase the number of CHLAC features, it is not a trivial matter to implement it; the method requires a large amount of computation time. This paper proposes a gait recognition method that achieves a higher recognition rate than the traditional CHLAC method. More specifically, our method has two distinguishing features. First, we have improved the accuracy of the traditional CHLAC method by increasing the number of CHLAC features. Second, we employ parallel computing techniques and a large-scale PC cluster to reduce the computational time. Libraries and tools used by the current implementation include MPICH [3, 4] (a communication library for parallel programming), GXP [5, 14] (a tool for remote job submission) and scalapack [13] (a library for parallel numerical calculation). We evaluated our method by applying it to a data set containing objects whose scales change dynamically. We ran the parallel version of the program on a 192-node cluster. The experimental results show that our method achieved a higher recognition rate than that of the traditional CHLAC method. The results also show that the parallel version of the program achieved a better speedup than the sequential version. Contributions of this paper are summarized as follows: We took the new approach (parallelizing the CHLAC method) to improve a recognition rate. We applied the cluster computing technologies to

2 Learning Phase Images Recognition Phase Images Differencing/Binarization Differencing/Binarization CHLAC Feature Extraction CHLAC Feature Extraction Linear Discriminant Analysis Recognition Processing Figure 1: Extraction of a second-order CHLAC feature Discriminant space Results the gait recognition, which to the authors knowledge, the technologies had not ever been applied to. Figure 2: Flow chart of recognition method with CHLAC features We achieved the impressive results (higher recognition rate as well as better speedup) with the parallel CHLAC method. The remainder of this paper is organized as follows. Section 2 presents an overview of CHLAC features. Section 3 explains the traditional recognition method with CHLAC features. Section 4 and Section 5 describe our recognition method and how the method is executed in parallel respectively. Section 6 presents performance measurements and discusses experimental results. The final section summarizes the paper. 2 CHLAC Features Cubic Higher-order Local AutoCorrelation (CHLAC) [7] is a temporal extension of Higherorder Local AutoCorrelation (HLAC) [11]. Specifically, CHLAC feature vector x = [x 1,..., x N ] t is defined as x i = f(r)f(r + a i 1)...f(r + a i D)dr, (1) where A A is cubic data which consists of successive image frames, r R 3 is a position vector of the reference point in A, a is a displacement vector, f(r) is the value at the point r, and D is the order of CHLAC. Figure 1 briefly illustrates how a second order CHLAC feature is extracted. We also define the reference point r as r = [x, y, t] t where x and y are the horizontal and vertical position in an image respectively, and t is the time. CHLAC extracts information about the shape and the motion of objects. They have some preferable characteristics for analysis of spatio-temporal data such as human Figure 3: An example of differenced and binarized image: moving points turn white (the other points stay black) motions. One is that they are model-free. In previous research, these features were applied to a gait recognition task and a high recognition rate was achieved. Another merit of CHLAC features is that they are shift-invariant. They absorb the difference of positions or the time lag. One of its merits is that its computational cost is constant regardless of the person in the image or his motion. 3 Traditional Recognition Method with CHLAC Features In this section, we explain each step of the traditional recognition method with CHLAC features. Figure 2 illustrates the flow chart of the method. The method has two phases: the learning phase and the recognition phase. Each phase consists of three steps. In the first step of learning phase, motion images are preprocessed. This step consists of image differencing and binarization. First, it creates difference images by calculating the differences between a pair of consecutive image frames. Then, it binarizes the difference images using an automatic threshold-selection technique [10]. This technique extracts only moving points from given images. An example of an extracted image is shown in Figure 3.

3 Then, CHLAC features are extracted from binarized images obtained in the previous step. In the traditional method, independent 251 kinds of CHLAC features are extracted from each image by limiting a displacement vector a in the equation (1) to a range of After that, linear discriminant analysis is applied to the extracted CHLAC feature vectors. In the recognition phase, CHLAC features of images from a video camera are extracted in the same way as learning data. Mahalanobis distances between an input CHLAC feature vector and the class mean vectors in the discriminant space are compared, and the input vector is classified to the nearest class. 4 Our Recognition Method While our recognition method basically follows the traditional method described in Section 3, our method has distinguishing features in CHLAC feature extraction, linear discriminant analysis, and recognition processing. This section describes the details of the differences below. Note that we employ N-fold cross validation to evaluate our recognition method. 4.1 CHLAC Feature Extraction In order to extract more CHLAC features from the binarized difference images, we take two different approaches. In the first approach, we rewrite the equation (1) using new parameters r x, r y and r t as follows: x i = f(r)f(r + Ra i 1)...f(r + Ra i D)dr (2) where A R = r x r y r t These parameters represent the range of displacement in each direction. We also vary the image resolution (the vertical/horizontal length) and the range of integration period (the number of image frames). By varying those parameters, we can obtain an effect similar to using multi-resolution images. We call a parameter set (including the number of image frames) a layer, and n parameter sets n layers. Since 251 CHLAC features are extracted from 1-layer, 251 n kinds of CHLAC features are obtained from n layers. In the second approach, we extend the range of the displacement vector from to As a result, 5527 kinds of CHLAC features are extracted, which are about 22 times as many as the traditional method (251). 4.2 Linear Discriminant Analysis and Recognition Process We adopt the different recognition process for the multi-layer approach and the displacement-vector extension approach. For the former approach, we adopt a majority method and a joint method. In the majority method, the final results of the recognition is decided by the majority of the results of the individual layers. The joint method obtains the final result by concatenating the CHLAC feature vector obtained from each layer into a single vector. For the latter approach, we use the same recognition process as that of the traditional method. More specifically, the majority method and the joint method perform recognition in the following ways: Majority method In the learning phase, for each layer, linear discriminant analysis is applied to extracted CHLAC feature vectors. With this analysis, we obtain information necessary for recognition: a mean vector and an inverse covariance matrix of each class in the discriminant space and discriminant axes. In the recognition phase, first, we project an extracted CHLAC feature vector to the discriminant space using the discriminant axes calculated in the learning phase. Next, we calculate the square of the Mahalanobis distances between each pair of the projected CHLAC feature vectors and class mean vectors. Then, we obtain the results of the recognition for each layer by classifying each vector to a class whose distance is the shortest. Finally, we decide the final result by a majority of the results of the individual layers. Joint method In the learning phase, linear discriminant analysis is applied to the single vector that concatenates extracted CHLAC feature vectors of the individual layers. In the recognition phase, we project combined CHLAC feature vectors to the discriminant space, and then we calculate the square of the Mahalanobis distances between each pair of the projected vector and class mean vectors. Then, we obtain the result of the recognition by classifying the vector to a class whose distance is the shortest. 5 Parallelization This section describes how we parallelize our recognition method in order to reduce the computational time. We parallelize time-consuming operations in both the learning phase and the recognition phase. Table 1 summarizes the target processes and parallelization techniques applied to the processes. In the rest of this section, we describe details of each parallelization. First, we explain how we parallelize the extraction of CHLAC features. Since CHLAC features can be extracted independently from each image data set, we can easily parallelize this step by equally-divided time-series image data to each processor (See Figure 4). We statically determines this allocation assuming that the feature extraction of each image requires almost same amount of computation time. In the current implementation, we achieve this parallelization using GXP [5, 14], a tool for automatically logging in many nodes and for submitting commands to them in parallel. Using the job-submission mechanism of GXP, we extract CHLAC features in parallel, collect the extracted CHLAC features, and store them as files on a hard-disk. More specifically, GXP provides fast parallel (simultaneous) command submission and parallel pipes (pipes between local command and all parallel commands). Combination of these mechanisms

4 Table 1: Parallelization of the learning phase and the recognition phase Description of target process Parallelization technique CHLAC feature extraction GXP Linear discriminant Calculation of covariance matrix MPICH analysis Calculation of eigenvalue problem scalapack N-fold cross validation GXP Image Sequences CHLAC Feature Vectors CHLAC Feature Vectors Covariance Matrix PC Cluster Figure 4: Parallelization of CHLAC feature extraction Figure 5: Parallelizing computation of a covariance matrix enables us to submit the programs to a large number of nodes in a cluster while giving different input parameters to the individual nodes. Second, we parallelize calculation of a covariance matrix required by linear discriminant analysis. We calculate covariance matrix of class Σ according to the following equation: Σ = 1 N N (x i x)(x i x) T i=1 where x i is a feature vector extracted from the ith data of the class, x is the average features of all the data of the class, and N is the number of data. To parallelize the calculation of the above equation, we assign equallydivided feature vectors to each processor, and let it compute (x i x)(x i x) T (Figure 5). Because this approach requires complex communication among processes and GXP is not suitable for such parallel processing, we implemented a parallel program for calculating the covariance matrix with MPICH [3, 4], which is a communication library for parallel programming. Third, we calculate an eigenvalue and eigenvector in parallel using scalapack [13], which is a library for parallel numerical calculation. We use the PDSYEVX routine, a solver for dense symmetric eigenproblem based on the parallel Householder reduction. Finally, we parallelize N-fold cross validation. Since individual cross validation can be executed in an independent manner, we perform every cross validation in parallel using GXP in the same way as feature extraction. Figure 6: An example of images used in the experiments 6 Gait Recognition Experiments 6.1 Experimental Environments We applied the recognition method described above to movements of people in which they go up and down stairs. Figure 6 shows an example of images used in these experiments. The recognition of such movements is a tough problem since the scale of the objects change dynamically. To our knowledge, there is no previous work targeting these movements. More specifically, we used data consisting of 11 people who moved in four different ways: (going up, going down) (walking, running). The people performed each movement six times, and the total number of the images was 44,966. For each data set, we tried to recognize the person who performed the movement and the kind of his/her

5 Table 2: Parameter range used for the recognition methods Recognition method Parameter range a. single layer with 251 features f {25, 30, 35, 40}, (w, h) {(360, 240), (240, 160), (180, 120), (120, 80)}, r x {1, 2, 3}, r y {1, 2, 3}, r t {1, 2, 3} b. majority with 3 layers any combination of 3 layers from (f, w, h, r x, r y, r t ) = (35, 360, 240, 3, 2, 1), (35, 240, 160, 2, 1, 1), (40, 360, 240, 2, 2, 1), (35, 240, 160, 1, 1, 1), (35, 240, 160, 2, 2, 1), (35, 240, 160, 3, 2, 1) (20 cases) c. majority with 3 layers any combination of 5 layers from (f, w, h, r x, r y, r t ) = (35, 360, 240, 3, 2, 1), (35, 240, 160, 2, 1, 1), (40, 360, 240, 2, 2, 1), (35, 240, 160, 1, 1, 1), (35, 240, 160, 2, 2, 1), (35, 240, 160, 3, 2, 1), (40, 360, 240, 3, 1, 1) (21 cases) d. majority with 7 layers any combination of 7 layers from (f, w, h, r x, r y, r t ) = (35, 360, 240, 3, 2, 1), (35, 240, 160, 2, 1, 1), (40, 360, 240, 2, 2, 1), (35, 240, 160, 1, 1, 1), (35, 240, 160, 2, 2, 1), (35, 240, 160, 3, 2, 1), (40, 360, 240, 3, 1, 1), (40, 360, 240, 3, 3, 1), (35, 240, 160, 3, 1, 1) (36 cases) e. joint with 3 layers same as majority with 3 layers f. joint with 5 layers same as majority with 5 layers g. joint with 7 layers same as as majority with 7 layers h. single layer with 5527 layers f = 40, (w, h) = (360, 240), r x = r y = r t = 1 f: range of integration period in equation (2) w, h: horizontal/vertical length of the image r x, r y, r t : range of displacement in each direction in equation (2) movement (11people 4 movements = 44 categories) with the following eight methods: the original version (single layer with 251 features) the majority method and joint method respectively with three, five, and seven layers the mono-layered method in which the dimension of displacement vector is extended to (single layer with 5527 features) For each recognition method, we searched the most suitable parameter set for recognition as shown in Table 2, and evaluated the recognition rates using 6-fold cross validation. The parallel version of the program was performed on a 192-node cluster with Xeon 2.4/2.8 GHz dual processors, while the sequential version on a Xeon 2.8 GHz processor. All the machines are equipped with 2 GB RAM and a 1 Gigabit Ethernet NIC. 6.2 Experimental Results Figure 7 summarizes the experimental results. Figure 7-(A) shows the highest recognition that each method achieved with varying parameters. These results show that our revised methods achieved up to about 97.7% recognition rate, while the original method (single layer with 251 features) achieved at most about 85% recognition rate. Figure 7-(B), (C), (D) and (E) show the execution time and the speedup rate of both feature extraction and cross validation. We can see that the total execution time of our recognition system was greatly reduced to a practical time due to the parallelization. For both feature extraction and cross validation step, all the methods achieved a speedup with parallelization. For example, recognition method h (single layer with 5527 features) took only minutes (about 3 hours) with parallelization while it took 7,140 minutes (about 5 days) without parallelization. For the feature extraction step, h achieved the lowest speedup ratio among all the methods while it achieved the highest for the cross validation step. We also note that the speed of the recognition processing (reading images + making differential/binarizing images + CHLAC feature extraction + recognition) on a Xeon 2.8 GHz processor was about 29 frames per second (fps) in the original method (a), and about 3.4 fps in the 7-layered joint method (g). 6.3 Discussion about the Recognition Rates We discuss three remarkable points about the recognition rates of the experiments (Figure 7-(A)). First, the recognition rate was improved as the number of features used for recognition increased. This result indicates that the extracted features become more expressive as the number of the features increases. Second, the joint methods generally have advantages in a recognition rate over the majority methods if both methods use the same number of CHLAC features. We can consider several reasons for this result. One possible reason is that CHLAC features are not scale-invariant. Since the data sets used in the experiments include objects that change their scale, 251 CHLAC features of the original method (a) were not sufficient for recognition. Another reason is that the current implementation does not support a mechanism that enables individual recognition systems to compensate for one another although it is a general technique for majority voting (e.g., boosting

6 (A) (B) (C) (D) (E) Figure 7: Experimental results: (A) recognition rate on the movements of going up/down stairs, (B) computation time of the feature extraction step, (C) speedup ratio in the feature extraction step, (D) computation time of the cross validation step, and (E) speedup ratio of the cross validation step techniques such as AdaBoost [2]). Third, the 7-layered joint method (g) achieved a 97.67% recognition rate, which is as high as that of the extended mono-layered method (h) even if they use the different dimension of the feature vectors (1757 and 5527 respectively). This result indicates that 5527 features in the extended mono-layered method (h) involve many meaningless features that do not represent the substance of the target objects. Since these meaningless features turn the result of discriminant analysis unstable, simply increasing the number of features does not necessarily improve the recognition rate. 7 Conclusion and Future Work We have presented a gait recognition method with CHLAC features. Our method achieves a higher recognition rate than the traditional one by increasing the number of CHLAC features. Since using a large number of CHLAC features requires a large amount of computation time, our method parallelizes the learning and recognition phase using a large-scale PC cluster. The experimental results show that our recognition method identifies the movements of going up/down stairs more accurately than the traditional one. It should be noted that our method can be easily applied to other kinds of movements (e.g., fighting action) because our method does not assume any model for target objects in advance. This is the advantage of our method over other model-based gait-recognition methods. We also note that there is room for further study about ways to increase the number of CHLAC features. As shown in equation 2, the current algorithm uses all kinds of displacement vectors within a certain range. However, irrelevant features should be excluded from the view point of recognition. Accordingly, we need to seek an algorithm that guesses in advance how each feature can contribute to recognition and that uses only meaningful features. Acknowledgements This work was performed as part of the Hundred- Hour Workshop at the Information Science and Technology Strategic Core, Graduate School of Information Science and Technology 21st Century COE Program, University of Tokyo. Details concerning this workshop can be found at WS100H.NET [1, 9]. References [1] A Hundred-Hour Workshop@UT-I-COE,. [2] Yoav Freund and Robert E. Schapire. A Decision- Theoretic Generalization of On-Line Learning and an Application to Boosting. Journal of Computer and System Sciences, 55(1): , [3] William Gropp and Ewing Lusk. User s Guide for MPICH, a Portable Implementation of MPI. Technical Report ANL-96/6, Argonne National Laboratory, 1994.

7 [4] William Gropp, Ewing Lusk, Nathan Doss, and Anthony Skjellum. A High-Performance, Portable Implementation of the MPI Message Passing Interface Standard. Parallel Computing, 22(6): , September [15] L. Wang, T. Tan, H. Ning, and W. Hu. Silhouette Analysis-Based Gait Recognition for Human Identification. IEEE Transactions on Pattern Analysis and Machine Intelligence, 25(12): , December [5] GXP Grid/Cluster Shell Quick Manual. phoenix/gxp_quick_man_ja.shtml. [6] A. Kale, N. Cuntoor, B. Yegnanarayana, A. N. Rajagopalan, and R. Chellappa. Gait Analysis for Human Identification. In Proceedings of International Conference on Audio- and Video-Based Biometric Person Authentication, [7] Takumi Kobayashi and Nobuyuki Otsu. Action and Simultaneous Multiple-Person Identification Using Cubic Higher Order Local Auto-Correlation. In Proceedings of 17th International Conference on Pattern Recognition (ICPR 2004), volume 4, pages , August [8] Takumi Kobayashi and Nobuyuki Otsu. A Three- Way Auto-Correlation Based Approach to Human Identification by Gait. In 6th IEEE Workshop on Visual Surveillance, pages , [9] Mihoko Otake, Ryo Fukano, Shinji Sako, Masao Sugi, Kiyoshi Kotani, Junya Hayashi, Hiroshi Noguchi, Ryuichi Yoneda, Kenjiro Taura, Nobuyuki Otsu, and Tomomasa Sato. Autonomous Collaborative Environment for Project Based Learning. In T. Arai et al., editor, Intelligent Autonomous Systems 9 (IAS-9), pages IOS Press, [10] Nobuyuki Otsu. A threshold selection method from gray-level histograms. IEEE Transactions on System, Man and Cybernetics, 9(1):62 66, March [11] Nobuyuki Otsu and Takio Kurita. A new scheme for practical, flexible and intelligent vision systems. In Proceedings of IAPR Workshop on Computer Vision Special Hardware and Industrical Applications, pages , October [12] Sudeep Sarkar, P. Jonathon Phillips, Zongyi Liu, Isidro Robledo Vega, Patrick Grother, and Kevin W. Bowyer. The HumanID Gait Challenge Problem: Data Sets, Performance, and Analysis. IEEE Transactions on Pattern Analysis and Machine Intelligence, 27(2): , February [13] scalapack. scalapack/. [14] Kenjiro Taura. GXP : An Interactive Shell for the Grid Environment. In Proceedings of International Workshop on Innovative Architecture for Future Generation High-Performance Processors and Systems (IWIA 2004), pages 59 67, January 2004.

Motion analysis for broadcast tennis video considering mutual interaction of players

Motion analysis for broadcast tennis video considering mutual interaction of players 14-10 MVA2011 IAPR Conference on Machine Vision Applications, June 13-15, 2011, Nara, JAPAN analysis for broadcast tennis video considering mutual interaction of players Naoto Maruyama, Kazuhiro Fukui

More information

Directionally-grouped CHLAC Motion Feature Extraction and Its Application to Sport Motion Analysis

Directionally-grouped CHLAC Motion Feature Extraction and Its Application to Sport Motion Analysis Directionally-grouped CHLAC Motion Feature Extraction and Its Application to Sport Motion Analysis FUMITO YOSHIKAWA, TAKUMI KOBAYASHI, KENJI WATANABE, NOBUYUKI OTSU National Institute of Advanced Industrial

More information

Similarity Image Retrieval System Using Hierarchical Classification

Similarity Image Retrieval System Using Hierarchical Classification Similarity Image Retrieval System Using Hierarchical Classification Experimental System on Mobile Internet with Cellular Phone Masahiro Tada 1, Toshikazu Kato 1, and Isao Shinohara 2 1 Department of Industrial

More information

Automatic Gait Recognition. - Karthik Sridharan

Automatic Gait Recognition. - Karthik Sridharan Automatic Gait Recognition - Karthik Sridharan Gait as a Biometric Gait A person s manner of walking Webster Definition It is a non-contact, unobtrusive, perceivable at a distance and hard to disguise

More information

Learning Human Identity using View-Invariant Multi-View Movement Representation

Learning Human Identity using View-Invariant Multi-View Movement Representation Learning Human Identity using View-Invariant Multi-View Movement Representation Alexandros Iosifidis, Anastasios Tefas, Nikolaos Nikolaidis and Ioannis Pitas Aristotle University of Thessaloniki Department

More information

Individual Recognition Using Gait Energy Image

Individual Recognition Using Gait Energy Image Individual Recognition Using Gait Energy Image Ju Han and Bir Bhanu Center for Research in Intelligent Systems University of California, Riverside, California 92521, USA jhan,bhanu @cris.ucr.edu Abstract

More information

Human Action Recognition Using Independent Component Analysis

Human Action Recognition Using Independent Component Analysis Human Action Recognition Using Independent Component Analysis Masaki Yamazaki, Yen-Wei Chen and Gang Xu Department of Media echnology Ritsumeikan University 1-1-1 Nojihigashi, Kusatsu, Shiga, 525-8577,

More information

Matching Gait Image Sequences in the Frequency Domain for Tracking People at a Distance

Matching Gait Image Sequences in the Frequency Domain for Tracking People at a Distance Matching Gait Image Sequences in the Frequency Domain for Tracking People at a Distance Ryusuke Sagawa, Yasushi Makihara, Tomio Echigo, and Yasushi Yagi Institute of Scientific and Industrial Research,

More information

Towards Understanding the Limits of Gait Recognition

Towards Understanding the Limits of Gait Recognition Towards Understanding the Limits of Gait Recognition Zongyi Liu, Laura Malave, Adebola Osuntugun, Preksha Sudhakar, and Sudeep Sarkar Computer Vision and Image Informatics Laboratory Computer Science and

More information

Recognition of Human Body Movements Trajectory Based on the Three-dimensional Depth Data

Recognition of Human Body Movements Trajectory Based on the Three-dimensional Depth Data Preprints of the 19th World Congress The International Federation of Automatic Control Recognition of Human Body s Trajectory Based on the Three-dimensional Depth Data Zheng Chang Qing Shen Xiaojuan Ban

More information

Extracting Spatio-temporal Local Features Considering Consecutiveness of Motions

Extracting Spatio-temporal Local Features Considering Consecutiveness of Motions Extracting Spatio-temporal Local Features Considering Consecutiveness of Motions Akitsugu Noguchi and Keiji Yanai Department of Computer Science, The University of Electro-Communications, 1-5-1 Chofugaoka,

More information

SPEECH FEATURE EXTRACTION USING WEIGHTED HIGHER-ORDER LOCAL AUTO-CORRELATION

SPEECH FEATURE EXTRACTION USING WEIGHTED HIGHER-ORDER LOCAL AUTO-CORRELATION Far East Journal of Electronics and Communications Volume 3, Number 2, 2009, Pages 125-140 Published Online: September 14, 2009 This paper is available online at http://www.pphmj.com 2009 Pushpa Publishing

More information

Mobile Face Recognization

Mobile Face Recognization Mobile Face Recognization CS4670 Final Project Cooper Bills and Jason Yosinski {csb88,jy495}@cornell.edu December 12, 2010 Abstract We created a mobile based system for detecting faces within a picture

More information

Extraction of Human Gait Features from Enhanced Human Silhouette Images

Extraction of Human Gait Features from Enhanced Human Silhouette Images 2009 IEEE International Conference on Signal and Image Processing Applications Extraction of Human Gait Features from Enhanced Human Silhouette Images Hu Ng #1, Wooi-Haw Tan *2, Hau-Lee Tong #3, Junaidi

More information

Face Recognition Technology Based On Image Processing Chen Xin, Yajuan Li, Zhimin Tian

Face Recognition Technology Based On Image Processing Chen Xin, Yajuan Li, Zhimin Tian 4th International Conference on Machinery, Materials and Computing Technology (ICMMCT 2016) Face Recognition Technology Based On Image Processing Chen Xin, Yajuan Li, Zhimin Tian Hebei Engineering and

More information

Text-Tracking Wearable Camera System for the Blind

Text-Tracking Wearable Camera System for the Blind 2009 10th International Conference on Document Analysis and Recognition Text-Tracking Wearable Camera System for the Blind Hideaki Goto Cyberscience Center Tohoku University, Japan hgot @ isc.tohoku.ac.jp

More information

Non-rigid body Object Tracking using Fuzzy Neural System based on Multiple ROIs and Adaptive Motion Frame Method

Non-rigid body Object Tracking using Fuzzy Neural System based on Multiple ROIs and Adaptive Motion Frame Method Proceedings of the 2009 IEEE International Conference on Systems, Man, and Cybernetics San Antonio, TX, USA - October 2009 Non-rigid body Object Tracking using Fuzzy Neural System based on Multiple ROIs

More information

Pixel-Pair Features Selection for Vehicle Tracking

Pixel-Pair Features Selection for Vehicle Tracking 2013 Second IAPR Asian Conference on Pattern Recognition Pixel-Pair Features Selection for Vehicle Tracking Zhibin Zhang, Xuezhen Li, Takio Kurita Graduate School of Engineering Hiroshima University Higashihiroshima,

More information

Video Aesthetic Quality Assessment by Temporal Integration of Photo- and Motion-Based Features. Wei-Ta Chu

Video Aesthetic Quality Assessment by Temporal Integration of Photo- and Motion-Based Features. Wei-Ta Chu 1 Video Aesthetic Quality Assessment by Temporal Integration of Photo- and Motion-Based Features Wei-Ta Chu H.-H. Yeh, C.-Y. Yang, M.-S. Lee, and C.-S. Chen, Video Aesthetic Quality Assessment by Temporal

More information

Uniprojective Features for Gait Recognition

Uniprojective Features for Gait Recognition Uniprojective Features for Gait Recognition Daoliang Tan, Kaiqi uang, Shiqi Yu, and Tieniu Tan Center for Biometrics and Security Research, National Laboratory of Pattern Recognition, Institute of Automation,

More information

P RW GEI: Poisson Random Walk based Gait Recognition

P RW GEI: Poisson Random Walk based Gait Recognition P RW GEI: Poisson Random Walk based Gait Recognition Pratheepan Yogarajah, Joan V. Condell and Girijesh Prasad Intelligent System Research Centre (ISRC) School of Computing and Intelligent Systems University

More information

Linear combinations of simple classifiers for the PASCAL challenge

Linear combinations of simple classifiers for the PASCAL challenge Linear combinations of simple classifiers for the PASCAL challenge Nik A. Melchior and David Lee 16 721 Advanced Perception The Robotics Institute Carnegie Mellon University Email: melchior@cmu.edu, dlee1@andrew.cmu.edu

More information

Video Inter-frame Forgery Identification Based on Optical Flow Consistency

Video Inter-frame Forgery Identification Based on Optical Flow Consistency Sensors & Transducers 24 by IFSA Publishing, S. L. http://www.sensorsportal.com Video Inter-frame Forgery Identification Based on Optical Flow Consistency Qi Wang, Zhaohong Li, Zhenzhen Zhang, Qinglong

More information

MULTI ORIENTATION PERFORMANCE OF FEATURE EXTRACTION FOR HUMAN HEAD RECOGNITION

MULTI ORIENTATION PERFORMANCE OF FEATURE EXTRACTION FOR HUMAN HEAD RECOGNITION MULTI ORIENTATION PERFORMANCE OF FEATURE EXTRACTION FOR HUMAN HEAD RECOGNITION Panca Mudjirahardjo, Rahmadwati, Nanang Sulistiyanto and R. Arief Setyawan Department of Electrical Engineering, Faculty of

More information

Parallel Architecture & Programing Models for Face Recognition

Parallel Architecture & Programing Models for Face Recognition Parallel Architecture & Programing Models for Face Recognition Submitted by Sagar Kukreja Computer Engineering Department Rochester Institute of Technology Agenda Introduction to face recognition Feature

More information

Face Recognition for Mobile Devices

Face Recognition for Mobile Devices Face Recognition for Mobile Devices Aditya Pabbaraju (adisrinu@umich.edu), Srujankumar Puchakayala (psrujan@umich.edu) INTRODUCTION Face recognition is an application used for identifying a person from

More information

Pattern Recognition Letters

Pattern Recognition Letters Pattern Recognition Letters xxx (2008) xxx xxx Contents lists available at ScienceDirect Pattern Recognition Letters journal homepage: www.elsevier.com/locate/patrec Three-way auto-correlation approach

More information

Fast Face Detection Assisted with Skin Color Detection

Fast Face Detection Assisted with Skin Color Detection IOSR Journal of Computer Engineering (IOSR-JCE) e-issn: 2278-0661,p-ISSN: 2278-8727, Volume 18, Issue 4, Ver. II (Jul.-Aug. 2016), PP 70-76 www.iosrjournals.org Fast Face Detection Assisted with Skin Color

More information

Expanding gait identification methods from straight to curved trajectories

Expanding gait identification methods from straight to curved trajectories Expanding gait identification methods from straight to curved trajectories Yumi Iwashita, Ryo Kurazume Kyushu University 744 Motooka Nishi-ku Fukuoka, Japan yumi@ieee.org Abstract Conventional methods

More information

Gait-based person identification method using shadow biometrics for robustness to changes in the walking direction

Gait-based person identification method using shadow biometrics for robustness to changes in the walking direction 2015 IEEE Winter Conference on Applications of Computer Vision Gait-based person identification method using shadow biometrics for robustness to changes in the walking direction Makoto Shinzaki 1, Yumi

More information

Moving Object Segmentation Method Based on Motion Information Classification by X-means and Spatial Region Segmentation

Moving Object Segmentation Method Based on Motion Information Classification by X-means and Spatial Region Segmentation IJCSNS International Journal of Computer Science and Network Security, VOL.13 No.11, November 2013 1 Moving Object Segmentation Method Based on Motion Information Classification by X-means and Spatial

More information

Use of Gait Energy Image in Implementation of Real Time Video Surveillance System

Use of Gait Energy Image in Implementation of Real Time Video Surveillance System IOSR Journal of Computer Engineering (IOSR-JCE) e-issn: 2278-0661, p- ISSN: 2278-8727Volume 16, Issue 1, Ver. 5 (Jan. 2014), PP 88-93 Use of Gait Energy Image in Implementation of Real Time Video Surveillance

More information

International Journal Of Global Innovations -Vol.6, Issue.I Paper Id: SP-V6-I1-P01 ISSN Online:

International Journal Of Global Innovations -Vol.6, Issue.I Paper Id: SP-V6-I1-P01 ISSN Online: IMPLEMENTATION OF OBJECT RECOGNITION USING SIFT ALGORITHM ON BEAGLE BOARD XM USING EMBEDDED LINUX #1 T.KRISHNA KUMAR -M. Tech Student, #2 G.SUDHAKAR - Assistant Professor, #3 R. MURALI, HOD - Assistant

More information

Distance-driven Fusion of Gait and Face for Human Identification in Video

Distance-driven Fusion of Gait and Face for Human Identification in Video X. Geng, L. Wang, M. Li, Q. Wu, K. Smith-Miles, Distance-Driven Fusion of Gait and Face for Human Identification in Video, Proceedings of Image and Vision Computing New Zealand 2007, pp. 19 24, Hamilton,

More information

An Adaptive Eigenshape Model

An Adaptive Eigenshape Model An Adaptive Eigenshape Model Adam Baumberg and David Hogg School of Computer Studies University of Leeds, Leeds LS2 9JT, U.K. amb@scs.leeds.ac.uk Abstract There has been a great deal of recent interest

More information

Change detection using joint intensity histogram

Change detection using joint intensity histogram Change detection using joint intensity histogram Yasuyo Kita National Institute of Advanced Industrial Science and Technology (AIST) Information Technology Research Institute AIST Tsukuba Central 2, 1-1-1

More information

HOG-Based Person Following and Autonomous Returning Using Generated Map by Mobile Robot Equipped with Camera and Laser Range Finder

HOG-Based Person Following and Autonomous Returning Using Generated Map by Mobile Robot Equipped with Camera and Laser Range Finder HOG-Based Person Following and Autonomous Returning Using Generated Map by Mobile Robot Equipped with Camera and Laser Range Finder Masashi Awai, Takahito Shimizu and Toru Kaneko Department of Mechanical

More information

FPGA Implementation of High Performance Face Detection Using MCT and Adaboost Algorithm

FPGA Implementation of High Performance Face Detection Using MCT and Adaboost Algorithm FPGA Implementation of High Performance Face Detection Using MCT and Adaboost Algorithm Ramya N P.G.Scholar, ECE, SKP Engineering College, Tiruvannamalai India E-mail : ramyamevlsi@gmail.com Srimathi Mathiazhagan

More information

Person identification from spatio-temporal 3D gait

Person identification from spatio-temporal 3D gait 200 International Conference on Emerging Security Technologies Person identification from spatio-temporal 3D gait Yumi Iwashita Ryosuke Baba Koichi Ogawara Ryo Kurazume Information Science and Electrical

More information

Human Motion Detection and Tracking for Video Surveillance

Human Motion Detection and Tracking for Video Surveillance Human Motion Detection and Tracking for Video Surveillance Prithviraj Banerjee and Somnath Sengupta Department of Electronics and Electrical Communication Engineering Indian Institute of Technology, Kharagpur,

More information

A Method of Annotation Extraction from Paper Documents Using Alignment Based on Local Arrangements of Feature Points

A Method of Annotation Extraction from Paper Documents Using Alignment Based on Local Arrangements of Feature Points A Method of Annotation Extraction from Paper Documents Using Alignment Based on Local Arrangements of Feature Points Tomohiro Nakai, Koichi Kise, Masakazu Iwamura Graduate School of Engineering, Osaka

More information

Monocular Vision-based Displacement Measurement System Robust to Angle and Distance Using Homography

Monocular Vision-based Displacement Measurement System Robust to Angle and Distance Using Homography 6 th International Conference on Advances in Experimental Structural Engineering 11 th International Workshop on Advanced Smart Materials and Smart Structures Technology August 1-2, 2015, University of

More information

A Novel Smoke Detection Method Using Support Vector Machine

A Novel Smoke Detection Method Using Support Vector Machine A Novel Smoke Detection Method Using Support Vector Machine Hidenori Maruta Information Media Center Nagasaki University, Japan 1-14 Bunkyo-machi, Nagasaki-shi Nagasaki, Japan Email: hmaruta@nagasaki-u.ac.jp

More information

Linear Discriminant Analysis in Ottoman Alphabet Character Recognition

Linear Discriminant Analysis in Ottoman Alphabet Character Recognition Linear Discriminant Analysis in Ottoman Alphabet Character Recognition ZEYNEB KURT, H. IREM TURKMEN, M. ELIF KARSLIGIL Department of Computer Engineering, Yildiz Technical University, 34349 Besiktas /

More information

ECE 172A: Introduction to Intelligent Systems: Machine Vision, Fall Midterm Examination

ECE 172A: Introduction to Intelligent Systems: Machine Vision, Fall Midterm Examination ECE 172A: Introduction to Intelligent Systems: Machine Vision, Fall 2008 October 29, 2008 Notes: Midterm Examination This is a closed book and closed notes examination. Please be precise and to the point.

More information

Detection of a Hand Holding a Cellular Phone Using Multiple Image Features

Detection of a Hand Holding a Cellular Phone Using Multiple Image Features Detection of a Hand Holding a Cellular Phone Using Multiple Image Features Hiroto Nagayoshi 1, Takashi Watanabe 1, Tatsuhiko Kagehiro 1, Hisao Ogata 2, Tsukasa Yasue 2,andHiroshiSako 1 1 Central Research

More information

RELIABILITY OF PARAMETRIC ERROR ON CALIBRATION OF CMM

RELIABILITY OF PARAMETRIC ERROR ON CALIBRATION OF CMM RELIABILITY OF PARAMETRIC ERROR ON CALIBRATION OF CMM M. Abbe 1, K. Takamasu 2 and S. Ozono 2 1 Mitutoyo Corporation, 1-2-1, Sakato, Takatsu, Kawasaki, 213-12, Japan 2 The University of Tokyo, 7-3-1, Hongo,

More information

Real-Time Document Image Retrieval for a 10 Million Pages Database with a Memory Efficient and Stability Improved LLAH

Real-Time Document Image Retrieval for a 10 Million Pages Database with a Memory Efficient and Stability Improved LLAH 2011 International Conference on Document Analysis and Recognition Real-Time Document Image Retrieval for a 10 Million Pages Database with a Memory Efficient and Stability Improved LLAH Kazutaka Takeda,

More information

FACE RECOGNITION BASED ON GENDER USING A MODIFIED METHOD OF 2D-LINEAR DISCRIMINANT ANALYSIS

FACE RECOGNITION BASED ON GENDER USING A MODIFIED METHOD OF 2D-LINEAR DISCRIMINANT ANALYSIS FACE RECOGNITION BASED ON GENDER USING A MODIFIED METHOD OF 2D-LINEAR DISCRIMINANT ANALYSIS 1 Fitri Damayanti, 2 Wahyudi Setiawan, 3 Sri Herawati, 4 Aeri Rachmad 1,2,3,4 Faculty of Engineering, University

More information

RECONSTRUCTION OF REGISTERED RANGE DATA USING GEODESIC DOME TYPE DATA STRUCTURE

RECONSTRUCTION OF REGISTERED RANGE DATA USING GEODESIC DOME TYPE DATA STRUCTURE RECONSTRUCTION OF REGISTERED RANGE DATA USING GEODESIC DOME TYPE DATA STRUCTURE Makoto Hirose and Kazuo Araki Graduate School of Computer and Cognitive Sciences, Chukyo University, Japan hirose@grad.sccs.chukyo-u.ac.jp,

More information

Face Recognition Using Vector Quantization Histogram and Support Vector Machine Classifier Rong-sheng LI, Fei-fei LEE *, Yan YAN and Qiu CHEN

Face Recognition Using Vector Quantization Histogram and Support Vector Machine Classifier Rong-sheng LI, Fei-fei LEE *, Yan YAN and Qiu CHEN 2016 International Conference on Artificial Intelligence: Techniques and Applications (AITA 2016) ISBN: 978-1-60595-389-2 Face Recognition Using Vector Quantization Histogram and Support Vector Machine

More information

A Pattern Matching Technique for Detecting Similar 3D Terrain Segments

A Pattern Matching Technique for Detecting Similar 3D Terrain Segments A Pattern Matching Technique for Detecting Similar 3D Terrain Segments MOTOFUMI T. SUZUKI National Institute of Multimedia Education 2-12 Wakaba, Mihama-ku, Chiba, 2610014 JAPAN http://www.nime.ac.jp/

More information

Parallel Architecture for Face Recognition using MPI

Parallel Architecture for Face Recognition using MPI Parallel Architecture for Face Recognition using MPI Dalia Shouman Ibrahim Computer Systems Department Computer and Information Sciences Ain Shams University Egypt Salma Hamdy Computer Science Department

More information

Automated Detection of Human for Visual Surveillance System

Automated Detection of Human for Visual Surveillance System Automated Detection of Human for Visual Surveillance System Yuji KUNO, Takahiro WATANABE, Yoshinori SHIMOSAKODA, and Satoshi NAKAGAWA Kansai Laboratory, Oki Electric Industry Co., Ltd., 2-27 Shiromi 1-chome,

More information

OCR For Handwritten Marathi Script

OCR For Handwritten Marathi Script International Journal of Scientific & Engineering Research Volume 3, Issue 8, August-2012 1 OCR For Handwritten Marathi Script Mrs.Vinaya. S. Tapkir 1, Mrs.Sushma.D.Shelke 2 1 Maharashtra Academy Of Engineering,

More information

A Fast and Accurate Eyelids and Eyelashes Detection Approach for Iris Segmentation

A Fast and Accurate Eyelids and Eyelashes Detection Approach for Iris Segmentation A Fast and Accurate Eyelids and Eyelashes Detection Approach for Iris Segmentation Walid Aydi, Lotfi Kamoun, Nouri Masmoudi Department of Electrical National Engineering School of Sfax Sfax University

More information

Motion in 2D image sequences

Motion in 2D image sequences Motion in 2D image sequences Definitely used in human vision Object detection and tracking Navigation and obstacle avoidance Analysis of actions or activities Segmentation and understanding of video sequences

More information

The Novel Approach for 3D Face Recognition Using Simple Preprocessing Method

The Novel Approach for 3D Face Recognition Using Simple Preprocessing Method The Novel Approach for 3D Face Recognition Using Simple Preprocessing Method Parvin Aminnejad 1, Ahmad Ayatollahi 2, Siamak Aminnejad 3, Reihaneh Asghari Abstract In this work, we presented a novel approach

More information

Gait Shape Estimation for Identification

Gait Shape Estimation for Identification Gait Shape Estimation for Identification David Tolliver and Robert T. Collins Robotics Institute, Carnegie Mellon University, Pittsburgh PA 15213, USA tolliver@ri.cmu.edu Abstract. A method is presented

More information

Efficient Block Matching Algorithm for Motion Estimation

Efficient Block Matching Algorithm for Motion Estimation Efficient Block Matching Algorithm for Motion Estimation Zong Chen International Science Inde Computer and Information Engineering waset.org/publication/1581 Abstract Motion estimation is a key problem

More information

Short Survey on Static Hand Gesture Recognition

Short Survey on Static Hand Gesture Recognition Short Survey on Static Hand Gesture Recognition Huu-Hung Huynh University of Science and Technology The University of Danang, Vietnam Duc-Hoang Vo University of Science and Technology The University of

More information

Automatic visual recognition for metro surveillance

Automatic visual recognition for metro surveillance Automatic visual recognition for metro surveillance F. Cupillard, M. Thonnat, F. Brémond Orion Research Group, INRIA, Sophia Antipolis, France Abstract We propose in this paper an approach for recognizing

More information

Gesture Recognition using Temporal Templates with disparity information

Gesture Recognition using Temporal Templates with disparity information 8- MVA7 IAPR Conference on Machine Vision Applications, May 6-8, 7, Tokyo, JAPAN Gesture Recognition using Temporal Templates with disparity information Kazunori Onoguchi and Masaaki Sato Hirosaki University

More information

Image Set-based Hand Shape Recognition Using Camera Selection Driven by Multi-class AdaBoosting

Image Set-based Hand Shape Recognition Using Camera Selection Driven by Multi-class AdaBoosting Image Set-based Hand Shape Recognition Using Camera Selection Driven by Multi-class AdaBoosting Yasuhiro Ohkawa, Chendra Hadi Suryanto, and Kazuhiro Fukui Graduate School of Systems and Information Engineering,

More information

A 3D Point Cloud Registration Algorithm based on Feature Points

A 3D Point Cloud Registration Algorithm based on Feature Points International Conference on Information Sciences, Machinery, Materials and Energy (ICISMME 2015) A 3D Point Cloud Registration Algorithm based on Feature Points Yi Ren 1, 2, a, Fucai Zhou 1, b 1 School

More information

Multidirectional 2DPCA Based Face Recognition System

Multidirectional 2DPCA Based Face Recognition System Multidirectional 2DPCA Based Face Recognition System Shilpi Soni 1, Raj Kumar Sahu 2 1 M.E. Scholar, Department of E&Tc Engg, CSIT, Durg 2 Associate Professor, Department of E&Tc Engg, CSIT, Durg Email:

More information

Person identification from human walking sequences using affine moment invariants

Person identification from human walking sequences using affine moment invariants 2009 IEEE International Conference on Robotics and Automation Kobe International Conference Center Kobe, Japan, May 2-7, 2009 Person identification from human walking sequences using affine moment invariants

More information

Chapter 4. The Classification of Species and Colors of Finished Wooden Parts Using RBFNs

Chapter 4. The Classification of Species and Colors of Finished Wooden Parts Using RBFNs Chapter 4. The Classification of Species and Colors of Finished Wooden Parts Using RBFNs 4.1 Introduction In Chapter 1, an introduction was given to the species and color classification problem of kitchen

More information

Evaluation of Hardware Oriented MRCoHOG using Logic Simulation

Evaluation of Hardware Oriented MRCoHOG using Logic Simulation Evaluation of Hardware Oriented MRCoHOG using Logic Simulation Yuta Yamasaki 1, Shiryu Ooe 1, Akihiro Suzuki 1, Kazuhiro Kuno 2, Hideo Yamada 2, Shuichi Enokida 3 and Hakaru Tamukoh 1 1 Graduate School

More information

AI Technology for Quickly Solving Urban Security Positioning Problems

AI Technology for Quickly Solving Urban Security Positioning Problems AI Technology for Quickly Solving Urban Security Positioning Problems Hiroaki Iwashita Kotaro Ohori Hirokazu Anai Security games are used for mathematically optimizing security measures aimed at minimizing

More information

Modeling Body Motion Posture Recognition Using 2D-Skeleton Angle Feature

Modeling Body Motion Posture Recognition Using 2D-Skeleton Angle Feature 2012 International Conference on Image, Vision and Computing (ICIVC 2012) IPCSIT vol. 50 (2012) (2012) IACSIT Press, Singapore DOI: 10.7763/IPCSIT.2012.V50.1 Modeling Body Motion Posture Recognition Using

More information

EE368 Project Report CD Cover Recognition Using Modified SIFT Algorithm

EE368 Project Report CD Cover Recognition Using Modified SIFT Algorithm EE368 Project Report CD Cover Recognition Using Modified SIFT Algorithm Group 1: Mina A. Makar Stanford University mamakar@stanford.edu Abstract In this report, we investigate the application of the Scale-Invariant

More information

Recognition of Non-symmetric Faces Using Principal Component Analysis

Recognition of Non-symmetric Faces Using Principal Component Analysis Recognition of Non-symmetric Faces Using Principal Component Analysis N. Krishnan Centre for Information Technology & Engineering Manonmaniam Sundaranar University, Tirunelveli-627012, India Krishnan17563@yahoo.com

More information

Accurate and Dense Wide-Baseline Stereo Matching Using SW-POC

Accurate and Dense Wide-Baseline Stereo Matching Using SW-POC Accurate and Dense Wide-Baseline Stereo Matching Using SW-POC Shuji Sakai, Koichi Ito, Takafumi Aoki Graduate School of Information Sciences, Tohoku University, Sendai, 980 8579, Japan Email: sakai@aoki.ecei.tohoku.ac.jp

More information

Production of Video Images by Computer Controlled Cameras and Its Application to TV Conference System

Production of Video Images by Computer Controlled Cameras and Its Application to TV Conference System Proc. of IEEE Conference on Computer Vision and Pattern Recognition, vol.2, II-131 II-137, Dec. 2001. Production of Video Images by Computer Controlled Cameras and Its Application to TV Conference System

More information

Matching Evaluation of 2D Laser Scan Points using Observed Probability in Unstable Measurement Environment

Matching Evaluation of 2D Laser Scan Points using Observed Probability in Unstable Measurement Environment Matching Evaluation of D Laser Scan Points using Observed Probability in Unstable Measurement Environment Taichi Yamada, and Akihisa Ohya Abstract In the real environment such as urban areas sidewalk,

More information

Time Stamp Detection and Recognition in Video Frames

Time Stamp Detection and Recognition in Video Frames Time Stamp Detection and Recognition in Video Frames Nongluk Covavisaruch and Chetsada Saengpanit Department of Computer Engineering, Chulalongkorn University, Bangkok 10330, Thailand E-mail: nongluk.c@chula.ac.th

More information

A Robust Method of Facial Feature Tracking for Moving Images

A Robust Method of Facial Feature Tracking for Moving Images A Robust Method of Facial Feature Tracking for Moving Images Yuka Nomura* Graduate School of Interdisciplinary Information Studies, The University of Tokyo Takayuki Itoh Graduate School of Humanitics and

More information

Photometric Stereo with Auto-Radiometric Calibration

Photometric Stereo with Auto-Radiometric Calibration Photometric Stereo with Auto-Radiometric Calibration Wiennat Mongkulmann Takahiro Okabe Yoichi Sato Institute of Industrial Science, The University of Tokyo {wiennat,takahiro,ysato} @iis.u-tokyo.ac.jp

More information

Motion Image Segmentation Using Global Criteria and DP

Motion Image Segmentation Using Global Criteria and DP Motion Image Segmentation Using Global Criteria and DP Takumi Kobayashi Fumito Yoshikawa Nobuyuki Otsu National Institute of Advanced Industrial Science and Technology Umezono 1-1-1, Tsukuba, Japan Abstract

More information

Performance Analysis of iscsi Middleware Optimized for Encryption Processing in a Long-Latency Environment

Performance Analysis of iscsi Middleware Optimized for Encryption Processing in a Long-Latency Environment Performance Analysis of iscsi Middleware Optimized for Encryption Processing in a Long-Latency Environment Kikuko Kamisaka Graduate School of Humanities and Sciences Ochanomizu University -1-1, Otsuka,

More information

Analyzing Dshield Logs Using Fully Automatic Cross-Associations

Analyzing Dshield Logs Using Fully Automatic Cross-Associations Analyzing Dshield Logs Using Fully Automatic Cross-Associations Anh Le 1 1 Donald Bren School of Information and Computer Sciences University of California, Irvine Irvine, CA, 92697, USA anh.le@uci.edu

More information

Data Reduction and Parallelization for Human Detection System

Data Reduction and Parallelization for Human Detection System Data Reduction and Parallelization for Human Detection System Mao Hatto, Takaaki Miyajima and Hideharu Amano Graduate School of Science and Technology, Keio University, Japan asap@am.ics.keio.ac.jp HOG

More information

Construction Progress Management and Interior Work Analysis Using Kinect 3D Image Sensors

Construction Progress Management and Interior Work Analysis Using Kinect 3D Image Sensors 33 rd International Symposium on Automation and Robotics in Construction (ISARC 2016) Construction Progress Management and Interior Work Analysis Using Kinect 3D Image Sensors Kosei Ishida 1 1 School of

More information

Measurement of Pedestrian Groups Using Subtraction Stereo

Measurement of Pedestrian Groups Using Subtraction Stereo Measurement of Pedestrian Groups Using Subtraction Stereo Kenji Terabayashi, Yuki Hashimoto, and Kazunori Umeda Chuo University / CREST, JST, 1-13-27 Kasuga, Bunkyo-ku, Tokyo 112-8551, Japan terabayashi@mech.chuo-u.ac.jp

More information

Semi-Supervised PCA-based Face Recognition Using Self-Training

Semi-Supervised PCA-based Face Recognition Using Self-Training Semi-Supervised PCA-based Face Recognition Using Self-Training Fabio Roli and Gian Luca Marcialis Dept. of Electrical and Electronic Engineering, University of Cagliari Piazza d Armi, 09123 Cagliari, Italy

More information

Face Detection and Recognition in an Image Sequence using Eigenedginess

Face Detection and Recognition in an Image Sequence using Eigenedginess Face Detection and Recognition in an Image Sequence using Eigenedginess B S Venkatesh, S Palanivel and B Yegnanarayana Department of Computer Science and Engineering. Indian Institute of Technology, Madras

More information

Mobile Phone Operations using Human Eyes Only and its Applications

Mobile Phone Operations using Human Eyes Only and its Applications Vol. 9, No. 3, 28 Mobile Phone Operations using Human Eyes Only and its Applications Kohei Arai Information Science Department Graduate School of Science and Engineering, Saga University Saga City, Japan

More information

Data Reduction and Parallelization for Human Detection System

Data Reduction and Parallelization for Human Detection System R2-4 SASIMI 2015 Proceedings Data Reduction and Parallelization for Human Detection System Mao Hatto, Takaaki Miyajima and Hideharu Amano Graduate School of Science and Technology, Keio University, Japan

More information

Fully Automatic Methodology for Human Action Recognition Incorporating Dynamic Information

Fully Automatic Methodology for Human Action Recognition Incorporating Dynamic Information Fully Automatic Methodology for Human Action Recognition Incorporating Dynamic Information Ana González, Marcos Ortega Hortas, and Manuel G. Penedo University of A Coruña, VARPA group, A Coruña 15071,

More information

CLASSIFICATION FOR ROADSIDE OBJECTS BASED ON SIMULATED LASER SCANNING

CLASSIFICATION FOR ROADSIDE OBJECTS BASED ON SIMULATED LASER SCANNING CLASSIFICATION FOR ROADSIDE OBJECTS BASED ON SIMULATED LASER SCANNING Kenta Fukano 1, and Hiroshi Masuda 2 1) Graduate student, Department of Intelligence Mechanical Engineering, The University of Electro-Communications,

More information

Mobile Human Detection Systems based on Sliding Windows Approach-A Review

Mobile Human Detection Systems based on Sliding Windows Approach-A Review Mobile Human Detection Systems based on Sliding Windows Approach-A Review Seminar: Mobile Human detection systems Njieutcheu Tassi cedrique Rovile Department of Computer Engineering University of Heidelberg

More information

Gait analysis for person recognition using principal component analysis and support vector machines

Gait analysis for person recognition using principal component analysis and support vector machines Gait analysis for person recognition using principal component analysis and support vector machines O V Strukova 1, LV Shiripova 1 and E V Myasnikov 1 1 Samara National Research University, Moskovskoe

More information

Face detection and recognition. Detection Recognition Sally

Face detection and recognition. Detection Recognition Sally Face detection and recognition Detection Recognition Sally Face detection & recognition Viola & Jones detector Available in open CV Face recognition Eigenfaces for face recognition Metric learning identification

More information

Tracking of Human Body using Multiple Predictors

Tracking of Human Body using Multiple Predictors Tracking of Human Body using Multiple Predictors Rui M Jesus 1, Arnaldo J Abrantes 1, and Jorge S Marques 2 1 Instituto Superior de Engenharia de Lisboa, Postfach 351-218317001, Rua Conselheiro Emído Navarro,

More information

A Novel Video Retrieval Method to Support a User's Recollection of Past Events Aiming for Wearable Information Playing

A Novel Video Retrieval Method to Support a User's Recollection of Past Events Aiming for Wearable Information Playing A Novel Video Retrieval Method to Support a User's Recollection of Past Events Aiming for Wearable Information Playing Tatsuyuki Kawamura, Yasuyuki Kono, and Masatsugu Kidode Graduate School of Information

More information

A Study on Similarity Computations in Template Matching Technique for Identity Verification

A Study on Similarity Computations in Template Matching Technique for Identity Verification A Study on Similarity Computations in Template Matching Technique for Identity Verification Lam, S. K., Yeong, C. Y., Yew, C. T., Chai, W. S., Suandi, S. A. Intelligent Biometric Group, School of Electrical

More information

Extracting Characters From Books Based On The OCR Technology

Extracting Characters From Books Based On The OCR Technology 2016 International Conference on Engineering and Advanced Technology (ICEAT-16) Extracting Characters From Books Based On The OCR Technology Mingkai Zhang1, a, Xiaoyi Bao1, b,xin Wang1, c, Jifeng Ding1,

More information

A Robust and Efficient Motion Segmentation Based on Orthogonal Projection Matrix of Shape Space

A Robust and Efficient Motion Segmentation Based on Orthogonal Projection Matrix of Shape Space A Robust and Efficient Motion Segmentation Based on Orthogonal Projection Matrix of Shape Space Naoyuki ICHIMURA Electrotechnical Laboratory 1-1-4, Umezono, Tsukuba Ibaraki, 35-8568 Japan ichimura@etl.go.jp

More information

A Minimum Number of Features with Full-Accuracy Iris Recognition

A Minimum Number of Features with Full-Accuracy Iris Recognition Vol. 6, No. 3, 205 A Minimum Number of Features with Full-Accuracy Iris Recognition Ibrahim E. Ziedan Dept. of computers and systems Faculty of Engineering Zagazig University Zagazig, Egypt Mira Magdy

More information