Video behavior recognition based on locally compressive sensing
WANG Wenfeng1, CHEN Xi1, WANG Haiyang2, PENG Wei3, QIAN Jing4, ZHENG Hongwei1
1. State Key Laboratory of Desert and Oasis Ecology, Xinjiang Institute of Ecology and Geography, Chinese Academy of Sciences, Urumqi 830011, China; 2. Institute of Computing Technology, Chinese Academy of Sciences, Beijing 100190, China; 3. School of Information Science and Engineering, Xiamen University, Xiamen 361005, China; 4. Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, Shenzhen 518055, China
Abstract：Compressive sensing has been successfully applied in the field of target tracking but not for behavior recognition. This paper presents a locally compressive sensing algorithm for behavior analysis which combines compressive tracking and centroid localization for recognition of video object behavior. Local compressive sensing selects a behavior-sensitive area for compressive tracking which characterizes target behavior based on classification of the object trajectory and local centroid velocity. Tests show that locally compressive sensing can accurately recognize global behavior such as running and falling and local behavior such as smiles and blinking. Therefore, the locally compressive sensing method is of great value that can be used for video surveillance and behavior recognition.
 KUMAR M, BHATNAGAR C. Crowd behavior recognition using hybrid tracking model and genetic algorithm enabled neural network[J]. International Journal of Computational Intelligence Systems, 2017, 10(1):234-246.  DEVANNE M, BERRETTI S, PALA P, et al. Motion segment decomposition of RGB-D sequences for human behavior understanding[J]. Pattern Recognition, 2017, 61:222-233.  WANG Y, CHEN H, LI S, et al. Object tracking by color distribution fields with adaptive hierarchical structure[J]. Visual Computer, 2017, 33(2):1-13.  CHEN Y, SHEN C. Performance analysis of smartphone-sensor behavior for human activity recognition[J]. IEEE Access, 2017, 5(3):3095-3110.  BATCHULUUN G, KIM J H, HONG H G, et al. Fuzzy system based human behavior recognition by combining behavior prediction and recognition[J]. Expert Systems with Applications, 2017, 81(9):108-133.  VAN V K, WASHINGTON G. Development of a wearable controller for gesture-recognition-based applications using polyvinylidene fluoride[J]. IEEE Transactions on Biomedical Circuits & Systems, 2017, 11(4):900-909.  ARABLOUEI R. Fast reconstruction algorithm for perturbed compressive sensing based on total least-squares and proximal splitting[J]. Signal Processing, 2017, 130(1):57-63.  DING X, CHEN W, WASSELL I J. Compressive sensing reconstruction for video:An adaptive approach based on motion estimation[J]. IEEE Transactions on Circuits & Systems for Video Technology, 2017, 27(7):1406-1420.  LAUE H E A. Demystifying compressive sensing[J]. IEEE Signal Processing Magazine, 2017, 34(4):171-176.  LIU T, QIU T, DAI R, et al. Nonlinear regression A*OMP for compressive sensing signal reconstruction[J]. Digital Signal Processing, 2017, 69:11-21.  JIANG H, DENG W, SHEN Z. Surveillance video processing using compressive sensing[J]. Inverse Problems & Imaging, 2017, 6(2):201-214.  KITAMURA T, IZUMI K, NAKAJIMA K, et al. Microlensed image centroid motions by an exotic lens object with negative convergence or negative mass[J]. Physical Review D, 2014, 89(8):1-2.  CAMPANA R, MASSARO E, BERNIERI E, et al. Application of the MST clustering to the high energy, γ-ray sky. I-New possible detection of high-energy, γ-ray emission associated with BL Lac objects[J]. Astrophysics and Space Science, 2015, 360(2):1-10.  MINGHU W U, ZHU X. Distributed video compressive sensing reconstruction by adaptive PCA sparse basis and nonlocal similarity[J]. Ksii Transactions on Internet & Information Systems, 2014, 8(8):2851-2865.  GU Y, GOODMAN N A. Information-theoretic compressive sensing kernel optimization and Bayesian Cramér-Rao bound for time delay estimation[J]. IEEE Transactions on Signal Processing, 2017, 65(17):4525-4537.  HEGDE C, INDYK P, SCHMIDT L. Approximation algorithms for model-based compressive sensing[J]. IEEE Transactions on Information Theory, 2015, 61(9):5129-5147.