This study is to propose robust approaches for pedestrian detection and pedestrian tracking, especially when there are occlusions for those targets. For pedestrian detection, a Gaussian Mixture Model (GMM) foreground mask is considered to increase the detection rate of pedestrians. Subsequently, the Haar-like and Local Binary Pattern (LBP) features are employed to define the body feature of human beings because those approaches have the invariance property for rotation and scaling. After defining those features, the Adaboost machine learning algorithm is further employed to train the features to form a robust detector in pedestrian detection procedure. For pedestrian tracking process, the AKAZE (accelerated-KAZE) is considered to compare the pedestrian Region-of-Interest (ROI) features from the detection process. Since the extracted ROI may have lower resolution, the Laplacian of Gaussian (LoG) is added into the AKAZE process to sharpen the features so as to obtain more feature key-points for matching. During the LoG-AKAZE matching process, a distance threshold is used to define the inlier and outlier key-points to increase the tracking accuracy. When occlusion occurred, the adaptive particle filter is triggered to predict the target position in the next frame. After that, the adaptive particle filter returns the position to the LoG-AKAZE to track continuously. Consequently, a robust surveillance system is built and can provide effectively pedestrian detection and tracking. Through the advantage of the proposed robust trained-detector and LoG-AKAZE, an intelligent surveillance system becomes possible and can possibly make the society more convenient and safe in the future.

This study is to propose robust approaches for pedestrian detection and pedestrian tracking, especially when there are occlusions for those targets. For pedestrian detection, a Gaussian Mixture Model (GMM) foreground mask is considered to increase the detection rate of pedestrians. Subsequently, the Haar-like and Local Binary Pattern (LBP) features are employed to define the body feature of human beings because those approaches have the invariance property for rotation and scaling. After defining those features, the Adaboost machine learning algorithm is further employed to train the features to form a robust detector in pedestrian detection procedure. For pedestrian tracking process, the AKAZE (accelerated-KAZE) is considered to compare the pedestrian Region-of-Interest (ROI) features from the detection process. Since the extracted ROI may have lower resolution, the Laplacian of Gaussian (LoG) is added into the AKAZE process to sharpen the features so as to obtain more feature key-points for matching. During the LoG-AKAZE matching process, a distance threshold is used to define the inlier and outlier key-points to increase the tracking accuracy. When occlusion occurred, the adaptive particle filter is triggered to predict the target position in the next frame. After that, the adaptive particle filter returns the position to the LoG-AKAZE to track continuously. Consequently, a robust surveillance system is built and can provide effectively pedestrian detection and tracking. Through the advantage of the proposed robust trained-detector and LoG-AKAZE, an intelligent surveillance system becomes possible and can possibly make the society more convenient and safe in the future.

[1] K. P. Chou et al., "Fast Deformable Model for Pedestrian Detection with Haar-like features," IEEE Symposium Series on Computational Intelligence (SSCI), Honolulu, pp. 1-8, 2017.
[2] Y. Can, B. Li, G. Xu, "Particle Filter Based Multi-pedestrian Tracking by HOG and HOF," 4th IEEE Int. Conf. on Information Science and Technology, pp. 714–717, 2014.
[3] R. Kalshaonkar and S. Kuwelkar, "Design of An Accurate Pedestrian Detection System Using Modified HOG and LSVM," International Conference on Computing, Communication and Automation (ICCCA), pp. 957-962, 2017.
[4] H. T. Niknejad, A. Takeuchi, S. Mita and D. McAllester, "On-road Multivehicle Tracking Using Deformable Object Model and Particle Filter With Improved Likelihood Estimation," IEEE Transactions on Intelligent Transportation Systems, vol. 13, no. 2, pp. 748-758, 2012.
[5] F. Li, F. You, R. Zhang, et al., "An Improved Real-time Detection and Localization Scheme for Pedestrian Based on Information Fusion," Int. J. Appl. Math. Stat., vol. 51, no. 22, pp. 99–107, 2013.
[6] P. Dollar, C. Wojek, B. Schiele, et al., "Pedestrian Detection: An Evaluation of the State of the Art," IEEE Trans. Pattern Anal. Mach. Intell., vol. 34, no. 4, pp. 743–761, 2012.
[7] R. Lienhart and J. Maydt, "An Extended Set of Haar-like Features for Rapid Object Detection," International Conference on Image Processing, vol. 1, pp. 03, 2002.
[8] J. Xu, Q. Wu, J. Zhang and Z. Tang, "Fast and Accurate Human Detection Using a Cascade of Boosted MS-LBP Features," IEEE Signal Processing Letters, vol. 19, no. 10, pp. 676-679, 2012.
[9] X. Wang, T. X. Han and S. Yan, "An HOG-LBP Human Detector with Partial Occlusion Handling," IEEE 12th International Conference on Computer Vision, pp. 32-39, 2009.
[10] W. Li, H. Ni, Y. Wang, B. Fu, P. Liu and S. Wang, "Detection of Partially Occluded Pedestrians by An Enhanced Cascade Detector," IET Intelligent Transport Systems, vol. 8, no. 7, pp. 621-630, 2014.
[11] A. Ranftl, F. Alonso-Fernandez, S. Karlsson and J. Bigun, "Real-time AdaBoost Cascade Face Tracker Based on Likelihood Map and Optical Flow," IET Biometrics, vol. 6, no. 6, pp. 468-477, 2017.
[12] D. Varga, L. Havasi and T. Szirányi, "Pedestrian Detection in Surveillance Videos Based on CS-LBP Feature," International Conference on Models and Technologies for Intelligent Transportation Systems (MT-ITS), pp. 413-417, 2015.
[13] C. P. Papageorgiou, M. Oren and T. Poggio, "A General Framework for Object Detection," Sixth International Conference on Computer Vision (IEEE Cat. No.98CH36271), pp. 555-562, 1998.
[14] L. Guo, P. S. Ge, Y. B. Zhao, M. H. Zhang, and L. H. Li, "Pedestrian Detection Based on HOG Features Optimized by Gentle AdaBoost in ROI," Journal of Convergence Information Technology, vol. 8, no. 2, 2013.
[15] N. Dalal and B. Triggs, "Histograms of Oriented Gradients for Human Detection," IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'05), vol. 1, pp. 86-893, 2005.
[16] K. Makino, T. Shibata, S. Yachida, T. Ogawa and K. Takahashi, " Moving-object Detection Method for Moving Cameras by Merging Background Subtraction and Optical Flow Methods," IEEE Global Conference on Signal and Information Processing (GlobalSIP) , pp. 383-387 , 2017.
[17] N. Friedman and S. Russell, "Image Segmentation in Video Sequences: A Probabilistic Approach," 13th Conf. Uncertainty in Artificial Intelligence, 1997.
[18] C. Stauffer, W. E. L. Grimson, "Adaptive Background Mixture Models for Real-time Tracking," IEEE Comput. Soc., vol. 2, 1999.
[19] H. Qiang, C. Qian and B. Zhong, "A Real-time Moving Target Tracking Algorithm Based on SIFT," International Conference on Security, Pattern Analysis, and Cybernetics (SPAC), pp. 569-572, 2017.
[20] H. J. Chien, C. C. Chuang, C. Y. Chen and R. Klette, "When To Use What Feature? SIFT, SURF, ORB, or A-KAZE Features for Monocular Visual Odometry," International Conference on Image and Vision Computing New Zealand (IVCNZ), pp. 1-6, 2016.
[21] T. M. Thanh, P. T. Hiep, T. M. Tam and K. Ryuji, "Frame-patch Matching Based Robust Video Watermarking Using KAZE Feature," IEEE International Conference on Multimedia and Expo (ICME) , pp. 1-6 , 2013.
[22] R. S. Babu, B. Radhakrishnan and L. P. Suresh, "Detection and extraction of roads from satellite images based on Laplacian of Gaussian operator," International Conference on Emerging Technological Trends (ICETT), pp. 1-7, 2016.
[23] Y. Du, Jing Tian, Linna San and Bingbing Yan, "Research on particle filter tracking algorithm based on multi-feature covariance matrix," International Conference on Control, Automation and Robotics (ICCAR), pp. 753-757, 2017.
[24] Y. H. Chen, H. Y. S. Lin and C. W. Su, "Full-Frame video stabilization via SIFT feature matching," International Conference on Intelligent Information Hiding and Multimedia Signal Processing, pp. 361-364, 2014.
[25] P. F. Alantarilla, J. Nuevo, and A. Bartoli, "Fast Explicit Diffusion for Accelerated Features in Nonlinear Scale Spaces," British Machine Vision Conference, p. 13.1-13.11, 2013.
[26] X. Lu and D. Li, "Research on Target Detection and Tracking System of Rescue Robot," Chinese Automation Congress (CAC), pp. 6623-6627, 2017.
[27] D. N. Pritha, L. Savitha and S. S. Shylaja, "Face Recognition by Feedforward Neural Network Using Laplacian of Gaussian Filter and Singular Value Decomposition," First International Conference on Integrated Intelligent Computing, pp. 56-61, 2010.
[28] Y. Liu, C. Lan, F. Yao, L. Li and C. Li, "Oblique Remote Sensing Image Matching Based on Improved AKAZE Algorithm," International Conference on Information Science and Technology (ICIST), pp. 448-454, 2016.
[29] Y. Guan, X. Chen, D. Yang and Y. Wu, "Multi-person Tracking-by-detection with Local Particle Filtering and Global Occlusion Handling," IEEE International Conference on Multimedia and Expo (ICME), pp. 1-6, 2014.
[23] H. Zhou, Y. Gao, G. Yuan and R. Ji, "Adaptive Multiple Cues Integration for Particle Filter Tracking," IET International Radar Conference, pp. 1-6, 2015.
[31] Z. Duan, Z. Cai and J. Yu, "Occlusion Detection and Recovery in Video Object Tracking Based on Adaptive Particle Filters," Chinese Control and Decision Conference, pp. 466-469 , 2009.
[32] Y. C. Liu, S. S. Huang, C. H. Lu, F. C. Chang and P. Y. Lin, "Thermal Pedestrian Detection Using Block LBP with Multi-level Classifier," International Conference on Applied System Innovation (ICASI), pp. 602-605, 2017.
[33] F. Li, R. Zhang and F. You, "Fast Pedestrian Detection and Dynamic Tracking for Intelligent Vehicles within V2V Cooperative Environment," IET Image Processing, vol. 11, no. 10, pp. 833-840, 2017.
[34] Y. Freund and R. E. Schapire, "A Decision-theoretic Generalization of On-line Learning and An Application to Boosting," Journal of Computer and System Sciences, vol. 55, pp 119-139, 1997
[35] H. Zhang, Y. Xie and C. Xu, "A Classifier Training Method for Face Detection Based on AdaBoost," International Conference on Transportation, Mechanical, and Electrical Engineering (TMEE), pp. 731-734, 2011.
[36] P. Alcantarilla, J. Nuevo, and A. Bartoli, "Fast Explicit Diffusion for Accelerated Features in Nonlinear Scale Spaces," British Machine Vision Conf., 2013.
[37] S. D. Pan, X. J. An, and H. G. He, “Optimal O Bilateral Filter with Arbitrary Spatial and Range Kernels Using Sparse Approximation,” Mathematical Problems in Engineering, vol. 2014, no. 1, pp. 1-11, 2014.
[38] L. Feng, Z. Wu, and X. Long, "Fast Image Diffusion for Feature Detection and Description," International Journal of Computer Theory and Engineering, vol. 8, no. 1. 2016.
[39] M. Talha and R. Stolkin, "Particle Filter Tracking of Camouflaged Targets by Adaptive Fusion of Thermal and Visible Spectra Camera Data," IEEE Sensors Journal, vol. 14, no. 1, pp. 159-166. 2014.