To solve the speed bottleneck of deformable part models in the detection process, this paper proposes a cascade deformable part model with rapid computation of feature pyramids for the detection process of the model. Because the speed of the detection is mainly determined by the two processes of the feature computation and the object location, a two-stage speedup algorithm is proposed. Firstly, sparsely-sampled feature pyramids on the scale are utilized to approximate finely-sampled multi-scale image features to speed up the process of feature computation. Then combined with the cascade algorithm in the location process, a sequence model is utilized to evaluate individual parts sequentially so as to rapidly prune most object hypotheses of small possibilities in order to speed up the process of object location. The experimental results on PASCAL VOC 2007 dataset and INRIA dataset show that the algorithm in the paper apparently speeds up the speed of detection with minor loss in detection precision.
李春伟,于洪涛,李邵梅,卜佑军. 一种基于可变形部件模型的快速对象检测算法[J]. 电子与信息学报, 2016, 38(11): 2864-2870.
LI Chunwei, YU Hongtao, LI Shaomei, BU Youjun. Rapid Object Detection Algorithm Based on Deformable Part Models. JEIT, 2016, 38(11): 2864-2870.
FELZENSZWALB P, GIRSHICK R, MCALLESTER D, et al. Object detection with discriminatively trained part based models[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2010, 32(9): 1627-1645. doi: 10.1109/TPAMI.2009.167.
[2]
YAO Benjamin, NIE Bruce, LIU Zicheng, et al. Animated pose templates for modeling and detecting human actions[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2014, 36(3): 436-452. doi: 10.1109/TPAMI. 2013.144.
[3]
WEN Jia, WANG Xueping, KONG Lingfu, et al. Using weighted part model for pedestrian detection in crowded scenes based on image segmentation[J]. Proceedings of the National Academy of Sciences, India Section A: Physical Scienes 2016, 86(1): 125-136. doi: 10.1007/s40010-015- 0231-3.
[4]
OROZCO J, MARTINEZ B, and PANTIC M. Empirical analysis of cascade deformable models for multi-view face detection[J]. Image and Vision Computing, 2015, 42(1): 47-61. doi: 10.1016/j.imavis.2015.07.002.
[5]
OHNBAR E and TRIVEDI M M. Learning to detect vehicles by clustering appearance patterns[J]. IEEE Transactions on Intelligent Transportation Systems, 2015, 16(5): 2511-2521. doi: 10.1109/TITS.2015.2409889.
[6]
DALAL N and TRIGGS B. Histograms of oriented gradients for human detection[C]. Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition, San Diego, USA, 2005: 886-893. doi: 10.1109/ CVPR.2005.177.
[7]
FELZENSZWALB P, GIRSHICK R, and MCALLESTER D. Cascade object detection with deformable part models[C]. Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition, San Francisco, USA, 2010: 2241-2248. doi: 10.1109/CVPR.2010.5539906.
[8]
PEDERSOLI M, VEDALDI A, GONZALEZ J, et al. A coarse-to-fine approach for fast deformable object detection[J]. Pattern Recognition, 2015, 48(7): 1844-1853. doi: 10.1016/j.patcog.2014.11.006.
[9]
ZHU Menglong, ATANASOV N, PAPPAS G J, et al. Active deformable part models inference[C]. Proceedings of the 13th European Conference on Computer Vision, Zurich, Switzerland, 2014: 281-296. doi: 10.1007/978-3-319- 10584-0_19.
[10]
KOKKINOS I. Bounding part scores for rapid detection with deformable part models[C]. Proceedings of the 12th European Conference on Computer Vision, Firenze, Italy, 2012: 41-50. doi: 10.1007/978-3-642-33885-4_5.
[11]
LIU Qi, HUANG Zi, and HU Fuqiao. Accelerating convolution-based detection model on GPU[C]. Proceedings of the IEEE Estimation, Detection and Information Fusion, Harbin, China, 2015: 61-66. doi: 10.1109/ICEDIF.2015. 7280163.
[12]
DUBOUT C and FLEURET F. Exact acceleration of linear object detectors[C]. Proceedings of the 12th European Conference on Computer Vision, Firenze, Italy, 2012: 301-311. doi: 10.1007/978-3-642-33712-3_22.
[13]
YAN Junjie, LEI Zhen, WEN Longyin, et al. The fastest deformable part model for object detection[C]. Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition, Columbus, USA, 2014: 2497-2504. doi: 10.1109/CVPR.2014.320.
[14]
SONG H O, GIRSHICK R, ZICKLER S, et al. Generalized sparselet models for real-time multiclass object recognition[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2015, 37(5): 1001-1012. doi: 10.1109/TPAMI. 2014.2353631.
[15]
PIRSIAVASH H. Steerable part models[C]. Proceedings of the 2012 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, Providence, USA, 2012: 3226-3233. doi: 10.1109/CVPR.2012.6248058.
[16]
KOKKINOS I. Shufflets: shared mid-level parts for fast object detection[C]. Proceedings of the 14th International Conference on Computer Vision, Sydney, Australia, 2013: 1393-1400. doi: 10.1109/ICCV.2013.176.
[17]
DEAN T, RUZON M, SEGAL M, et al. Fast, accurate detection of 100,000 object classes on a single machine[C]. Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition, Portland, USA, 2013: 1814-1821. doi: 10.1109/CVPR.2013.237.
[18]
RUDERM D L. The statistics of natural images[J]. Network Computation in Neural Systems, 2009, 5(4): 517-548. doi: 10.1088/0954-898X_5_4_006.
[19]
DOLLAR P, APPEL R, BELONGIE S, et al. Fast feature pyramids for object detection[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2014, 36(8): 1532-1545. doi: 10.1109/TPAMI.2014.2300479.
[20]
HOSANG J, BENENSON R, DOLLAR P, et al. What makes for effective detection proposals?[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2016, 38(4): 814-830. doi: 10.1109/TPAMI.2015.2465908.