Target Tracking Based on Multiple Instance Deep Learning
Cheng Shuai① Sun Jun-xi② Cao Yong-gang①③ Liu Guang-wen① Hang Guang-liang③
①(School of Electronic Information Engineering, Changchun University of Science and Technology, Changchun 130022, China) ②(School of Computer Science and Information Technology, Northeast Normal University, Changchun 130117, China) ③(Changchun Institute of Optics, Fine Mechanics and Physics, Chinese Academy of Sciences, Changchun 130000, China)
To overcome the problem that the deficiency of the appearance model and the motion model often leads to low precision in original Multiple Instance Learning (MIL), a target tracking algorithm is proposed based on multiple instance deep learning. In original MIL algorithm, the image is not represented effectively by Haar-like feature. To improve the tracking precision, a stacked denoising autoencoder is used to learn image features and express the image representations obtained effectively. Selected feature vector could not be replaced in the original MIL algorithm, which has difficulty reflecting the changes of the target and the background．Thus, some weakest discriminative feature vector is replaced with new randomly generated feature vector when weak classifiers are selected. It introduces new information to the target model and adapts to the dynamic changes of the target. Aiming at the deficiency of using motion model where the location of the target is likely to appear within a radius in original MIL algorithm, the particle filter estimates object’s location to increase the tracking precision. Compared with the original MIL algorithm and other state-of-the-art trackers in the complex environment, the experiments on variant image sequences show that the proposed algorithm raise the tracking accuracy and the robustness.
Wu Y, Lim J, and Yang M H. Online object tracking: a benchmark[C]. Proceedings of IEEE Conference on Computer Vision and Pattern Recognition, Portland, USA, 2013: 2411-2418.
[2]
Ross D A, Lim J, Lin R S, et al.. Incremental learning for robust visual tracking[J]. International Journal of Computer Vision, 2008, 77(3): 125-141.
[3]
Zhang K, Zhang L, and Yang M H. Fast compressive tracking[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2014, 36(10): 2002-2015.
[4]
Zhong W, Lu H C, and Yang M H. Robust object tracking via sparse collaborative appearance model[J]. IEEE Transactions on Image Processing, 2014, 23(5): 2356-2368.
Chen S, Su S Z, Li S Z, et al.. A novel co-training object tracking algorithm based on online semi-supervised boosting[J]. Journal of Electronics & Information Technology, 2014, 36(4): 888-895.
[6]
Zhang K, Zhang L, Liu Q, et al.. Fast tracking via dense spatio-temporal context learning[C]. Proceedings of European Conference on Computer Vision, Zurich, Switzerland, 2014: 127-141.
[7]
Kalal Z, Mikolajczyk K, and Matas J. Tracking- learning-detection[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2012, 34(7): 1409-1422.
[8]
Babenko B, Yang M H, and Belongie S. Robust object tracking with online multiple instance learning[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2011, 33(8): 1619-1632.
[9]
Zhang K H and Song H H. Real-time visual tracking via online weighted multiple instance learning[J]. Pattern Recognition, 2013, 46(1): 397-411.
Chen D C, Zhu M, Gao W, et al.. Real-time object tracking via online weighted multiple instance learning[J]. Optics and Precision Engineer, 2014, 22(6): 1661-1667.
Ning J F, Zhao Y B, and Shi W Z. Multiple instance learning based object tracking with multi-channel haar-like feature[J]. Journal of Image and Graphics, 2014, 19(7): 1038-1045.
Zheng Y, Chen Q, and Zhang Y. Deep learning and its new progress in object and behavior recognition[J]. Journal of Image and Graphics, 2014, 19(2): 175-184.
[13]
Vincent P, Larochellel H, Lajoie I, et al.. Stacked denoising autoencoders: learning useful representations in a deep network with a local denoising criterion[J]. Journal of Machine Learning Research, 2010, 11: 3371-3408.
Cheng S, Cao Y G, Sun J X, et al.. Target tracking based on enhanced flock of tracker and deep learning[J]. Journal of Electronics & Information Technology, 2015, 37(7): 1646-1653.
[15]
Hinton G E and Salakhutdinov R R. Reducing the dimensionality of data with neural networks[J]. Science, 2006, 313(5786): 504-507.
[16]
Rob H and Alan F. Discriminatively trained particle filters for complex multi-object tracking[C]. Proceedings of IEEE Conference on Computer Vision and Pattern Recognition, Miami, USA, 2009: 240-247.
Li T C and Sun S D. Double-resampling based monte carlo localization for mobile robot[J]. Acta Automatica Sinica, 2010, 36(9): 1279-1286.
[18]
Torralba A, Fergus R, and Freeman W T. 80 million tiny images: a large data set for nonparametric object and scene recognition[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2008, 30(11): 1958-1970.
[19]
Olshausen B and Field D. Sparse coding with an overcomplete basis set: a strategy employed by V1[J]. Vision Research, 1997, 37(23): 3311-3326.
[20]
Dinh T B, Vo N, and Medion G. Context tracker: exploring supporters and distracters in unconstrained environments[C]. Proceedings of IEEE Conference on Computer Vision and Pattern Recognition, Providence, USA, 2011: 1177-1184.
[21]
Zhang T, Ghanem B, Liu S, et al.. Robust visual tracking via multi-task sparse learning[C]. Proceedings of IEEE Conference on Computer Vision and Pattern Recognition, Providence, USA, 2012: 2042-2049.
