Gait recognition based on Wasserstein generating adversarial image inpainting network

Abstract

Aiming at the problem of small area human occlusion in gait recognition, a method based on generating adversarial image inpainting network was proposed which can generate a context consistent image for gait occlusion area. In order to reduce the effect of noise on feature extraction, the stacked automatic encoder with robustness was used. In order to improve the ability of gait classification, the sparse coding was used to express and classify the gait features. Experiments results showed the effectiveness of the proposed method in comparison with other state-of-the-art methods on the public databases CASIA-B and TUM-Gaid for gait recognition.

Keywords

gait recognition / image inpainting / generating adversarial network / stacking automatic encoder

References

[1] DengM, WangC, ChengF. Fusion of spatial-temporal and kinematic features for gait recognition with deterministic learning [J]. Pattern Recognition, 2017, 67: 186-200

[2] LuW, ZongW, XingW, BaoE. Gait recognition based on joint distribution of motion angles [J]. Journal of Visual Languages & Computing, 2014, 25(6): 754-763

[3] FrancescoB, AlfredoP. TGLSTM: A time based graph deep learning approach to gait recognition [J]. Pattern Recognition letters, 2019, 126(4): 132-138

[4] SunB L, ZhangZ, LiuX Y, HuB, ZhuT S. Self-esteem recognition based on gait pattern using Kinect [J]. Gait & Posture, 2017, 58(3): 428-432

[5] YangK, DouY, LvS, ZhangF, LvQ. Relative distance features for gait recognition with Kinect [J]. Journal of Visual Communication and Image Representation, 2016, 39: 209-217

[6] López-FernándezD, Madrid-CuevasF J, Carmona-PoyatoA, Muñoz-SalinaS, Medina-CarnicerR. A new approach for multi-view gait recognition on unconstrained paths [J]. Journal of Visual Communication and Image Representation, 2016, 38(7): 396-406

[7] LuoJ, TangJ, TjahjadiT, XiaoX. Robust arbitrary view gait recognition based on parametric 3D human body reconstruction and virtual posture synthesis [J]. Pattern Recognition, 2016, 60: 361-377

[8] LiW, KuoJ, PengJ. Gait recognition via GEI subspace projections and collaborative representation classification [J]. Neurocomputing, 2018, 275: 1932-1945

[9] Das, ChoudhuryS, TjahjadiT. Clothing and carrying condition invariant gait recognition based on rotation forest [J]. Pattern Recognition Letters, 2016, 80: 1-7

[10] AmerA T, KhaledA, ShanablehT. Decision-level fusion for single-view gait recognition with various carrying and clothing conditions [J]. Image and Vision Computing, 2017, 61: 54-69

[11] YuS, ChenH, WangQ, ShenL, HuangY. Invariant feature extraction for gait recognition using only one uniform model [J]. Neurocomputing, 2017, 239: 81-93

[12] WuZ, HuangY, WangL. A comprehensive study on cross-view gait based human identification with deep CNNs [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2017, 39(2): 209-226

[13] PratikC, ShamikS, JayantaM. Frontal gait recognition from occluded scenes [J]. Pattern Recognition Letters, 2015, 63: 9-15

[14] XieJ, XuL, ChenEImage denoising and inpainting with deep neural networks [C]// 26th Annual Conference on Neural Information Processing Systems 2012, 2012, IEEE, Lake Tahoe, Nevada: 341349

[15] PathakD, KrahenbuhlP, DonahueJ, DarrellT, EfrosA AContext encoders: feature learning by inpainting [C]// 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2016, Las Vegas, USA, IEEE: 25362544

[16] YehR A, ChenC, LimT Y, SchwingA G, Hasegawa-JohnsonM, DoM NSemantic Image Inpainting with Deep Generative Models [C]// 2017 IEEE Conference on Computer Vision and Pattern Recognition, 2017, IEEE, Honolulu, HI: 68826890

[17] GoodfellowI J, PougetA J, MirzaMGenerative ad-versarial nets [C]// Proceedings of the 26th Annual Conference on Neural Information Processing Systems, 2014, IEEE, Montreal, Canada: 26722680

[18] ArjovskyM, ChintalaS, BottouLWasserstein GAN [EB/OL], 2018

[19] LuoJ, XuY, TangC, LvJLearning inverse mapping by autoencoder based generative adversarial nets [C]// International Conference on Neural Information Processing, 2017, Germany, Springer: 207216

[20] MohammadA K, DennisS M, MinhoL. Coupled generative adversarial stacked auto-encoder: CoGASA [J]. Neural Networks, 2018, 100: 1-9

[21] HintonG, SrivastavaN, SwerskyKNeural networks for machine learning: overview of mini-batch gradient descent [EB/OL], 2018

[22] WangJ, ZhouS-c, XiaL-min. Human interaction recognition based on sparse representation of feature covariance matrices [J]. Journal of Central South University, 2018, 25(2): 304-314

[23] AharonM, EladM, BrucksteinA. k-SVD: An algorithm for designing overcomplete dictionaries for sparse representation [J]. IEEE Transactions on Signal Processing, 2006, 54(11): 4311-4322

[24] LiY, LiuS, YangJ, YangMGenerative face completion [C]// 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2017, Honolulu, USA, IEEE: 58925900

[25] LishaniA O, BoubchirL, KhalifaE, BouridaneA. Human gait recognition based on Haralick features [J]. Signal Image and Video Processing, 2017, 11(6): 1123-1130

[26] AlotaibiM, MahmoodA. Improved Gait recognition based on specialized deep convolutional neural networks [J]. Computer Vision and Image Understanding, 2017, 164: 103-110

[27] KarenS, AndrewZVery deep convolutional networks for large-scale image recognition [C]// 3rd International Conference on Learning Representations, 2015, San Diego, ICLR: 178184

[28] ShiD KGait recognition algorithm research and application platform design based on human geometric characteristics [D], 2017, Zhejiang, Zhejiang University

[29] ZhangL, ZhangL, TaoD, DuB. A sparse and discriminative tensor to vector projection for human gait feature representation [J]. Signal Processing, 2015, 106(C): 245

[30] RidaI, MaadeedS A, BouridaneAUnsupervised feature selection method for improved human gait recognition [C]// 2015 23rd European Signal Processing Conference (EUSIPCO), 2015, Nice, New York, USA, IEEE: 1128-1132

[31] ChenX, XuJ. Uncooperative gait recognition: Re-ranking based on sparse coding and multi-view hypergraph learning [J]. Pattern Recognition, 2016, 53: 116-129