AUTHORS: Metzli Ramirez-Martinez, Francisco Sanchez-Fernandez, Philippe Brunet, Sidi-Mohammed Senouci, El-Bay Bourennane

Download as PDF

ABSTRACT: This work proposes and implements a method based on Context-Aware Visual Attention Model (CAVAM), but modifying the method in such way that the detection algorithm is replaced by Histograms of Oriented Gradients (HOG). After reviewing different algorithms for people detection, we select HOG method because it is a very well known algorithm, which is used as a reference in virtually all current research studies about automatic detection. In addition, it produces accurate results in significantly less time than many algorithms. In this way, we show that CAVAM model can be adapted to other methods for object detection besides Scale-Invariant Feature Transform (SIFT), as it was originally proposed. Additionally, we use TUD dataset image sequences to evaluate and compare our approach with the original HOG algorithm. These experiments show that our method achieves around 2x speed-up at just 2% decreased accuracy. Moreover, the proposed approach can improve precision and specificity by more than 2%.

KEYWORDS: Object detection, pedestrian detection, tile-based method, saliency, regions of interest

REFERENCES:

[1] N. Dalal and B. Triggs, Histograms of Oriented Gradients for Human Detection, IEEE Trans. Computer Society Conference on Computer Vision and Pattern Recognition (CVPR’05), 1, 886–893, 2005.

[2] M. Riesenhuber and T. Poggio, Hierarchical models of object recognition in cortex, Nature neuroscience 2(11), 1019–25, 1999.

[3] D. G. Lowe, Sift, in Computer vision. The proceedings of the seventh IEEE international conference on, 2, 1150–1157, 1999.

[4] H. Bay, T. Tuytelaars, and L. Van Gool, SURF: Speeded up robust features, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) 3951 LNCS(Chapter 32), 404–417, 2006.

[5] M. Szarvas, a. Yoshizawa, M. Yamamoto, et al., Pedestrian detection with convolutional neural networks, IEEE Proceedings. Intelligent Vehicles Symposium, 224–229, 2005.

[6] J. Oh, G. Kim, J. Park, et al., A 320 mW 342 GOPS real-time dynamic object recognition processor for HD 720p video streams, IEEE Journal of Solid-State Circuits 48(1), 33–45, 2013.

[7] S. Zhang, D. A. Klein, C. Bauckhage, et al., Fast moving pedestrian detection based on motion segmentation and new motion features, Multimedia Tools and Applications 75, 6263–6282, 2016.

[8] S. Tang, M. Andriluka, and B. Schiele, Detection and tracking of occluded people, International Journal of Computer Vision 110(1), 58–69, 2014.

[9] P. Felzenszwalb, D. McAllester, and D. Ramanan, A discriminatively trained, multiscale, deformable part model, 26th IEEE Conference on Computer Vision and Pattern Recognition, CVPR , 1–8, 2008.

[10] W. Li, H. Su, F. Pan, et al., A fast pedestrian detection via modified HOG feature, Chinese Control Conference, CCC 2015-Septe, 3870–3873, 2015.

[11] S. Leutenegger, M. Chli, and R. Siegwart, fBRISKg: Binary Robust Invariance Scalable Keypoints, Proceedings of the International Conference on Computer Vision (fICCVg) , 2548–2555, 2011.

[12] A. Alahi, R. Ortiz, and P. Vandergheynst, fFREAKg: Fast Retina Keypoint, fIEEEg Int. Conf. Computer Vision and Pattern Recognition (CVPR) , 510–517, 2012.

[13] C. Schaeffer, A Comparison of Keypoint Descriptors in the Context of Pedestrian Detection: FREAK vs. SURF vs. BRISK, tech. rep., Stanford University CS Department, California, 2012.

[14] L. Wang and B. Zhang, Boosting-Like Deep Learning For Pedestrian Detection, arXiv, 2015.

[15] X. Chen, P. Wei, W. Ke, et al., Pedestrian Detection with Deep Convolutional Neural Network, 354–365, 2015.

[16] S. Tang, M. Ye, C. Zhu, et al., Adaptive pedestrian detection using convolutional neural network with dynamically adjusted classifier, Journal of Electronic Imaging 26(1), 013012, 2017.

[17] A. Suleiman, Y. H. Chen, J. Emer, et al., Towards closing the energy gap between HOG and CNN features for embedded vision, Proceedings - IEEE International Symposium on Circuits and Systems, 2017.

[18] T. Nguyen, E.-a. Park, J. Han, et al., Object Detection Using Scale Invariant Feature Transform, Genetic and Evolutionary Computing , 65–72, 2014.

[19] R. E. Kalman, A New Approach to Linear Filtering and Prediction Problems 1, Journal of Fluids Engineering 82, 35–45, 1960.

[20] F. Perazzi, P. Krahenbuhl, Y. Pritch, et al., Saliency Filters : Contrast Based Filtering for Salient Region Detection, IEEE Trans. Computer Vision and Pattern Recognition (CVPR), 733–740, 2012.

[21] M.-m. Cheng, N. J. Mitra, X. Huang, et al., Global Contrast based Salient Region Detection, IEEE Transactions on Pattern Analysis and Machine Intelligence 37(3), 569–582, 2015.

[22] S. Goferman, L. Zelnik-Manor, and A. Tal, Context-Aware Saliency Detection, IEEE Transactions on Pattern Analysis and Machine Intelligence 34(10), 1915 – 1926, 2012.

[23] T. Liu, Z. Yuan, J. Sun, et al., Learning to Detect a Salient Object, IEEE Transactions on Pattern Analysis and Machine Intelligence 33(2), 353–367, 2011.

[24] L. Itti, C. Koch, and E. Niebur, A Model of Saliency-Based Visual Attention for Rapid Scene Analysis, IEEE Transactions on Automatic Control 20(11), 1254–1259, 1998.

[25] E. D. Gelasca, D. Tomasic, and T. Ebrahimi, Which colors best catch your eyes: a subjective study of color saliency, First International Workshop on Video Processing and Quality Metrics for Consumer Electronics VPQM-05 , 16, 2005.

[26] S. Lee, K. Kim, J. Y. Kim, et al., Familiarity based unified visual attention model for fast and robust object recognition, Pattern Recognition 43(3), 1116–1128, 2010.

[27] M. Andriluka, S. Roth, and B. Schiele, Peopletracking-by-detection and people-detectionbytracking, IEEE Conference on Computer Vision and Pattern Recognition, 1–8, 2008.

[28] C. Wojek, S. Walk, and B. Schiele, Multi-cue onboard pedestrian detection, in Computer Vision and Pattern Recognition, CVPR, 794–801, 2009.

[29] M. Andriluka, S. Roth, and B. Schiele, Pictorial structures revisited: People detection and articulated pose estimation, IEEE Conference on Computer Vision and Pattern Recognition, CVPR, 1014–1021, 2009.

[30] G. Overett, L. Petersson, N. Brewer, et al., A new pedestrian dataset for supervised learning, IEEE Trans. Intelligent Vehicles Symposium, 373–378, 2008.