Key points and visible part fusion attention network for occluded pedestrian detection in traffic environments

Abstract

Aiming at the problem of low detection accuracy of occluded pedestrian in traffic environments, this paper proposes a key points and visible part fusion network for occluded pedestrian detection. The proposed algorithm constructs two attention modules by introducing human key points and the bounding box of visible parts respectively, which suppresses the occluded parts in the channel features and spatial features of pedestrian features respectively. Experimental results on CityPersons and Caltech datasets demonstrate the effectiveness of the proposed algorithm. The missing rate (MR) is reduced to 40.78 on the Heavy subset of the CityPersons dataset and surpasses many outstanding methods.

References

[[1]] Liu M Y, Jiang J, Zhu C, et al.. . VLPD: context-aware pedestrian detection via vision-language semantic self-supervision[C], 2023 New York IEEE 6662-6671

[[2]] Abdul K, Mohammed N, Andreas D. . Localized semantic feature mixers for efficient pedestrian detection in autonomous driving[C], 2023 New York IEEE 5476-5485

[[3]] Tang Y, Liu M, Li B P, et al.. OTP-NMS: toward optimal threshold prediction of NMS for crowded pedestrian detection[J]. IEEE transactions on image processing, 2023, 32: 3176-3187,

[[4]] Sundararaman R, De A C, Marchand E, et al.. . Tracking pedestrian heads in dense crowd[C], 2021 New York IEEE 3865-3875

[[5]] Liu L, Chen J, Wu H, et al.. . Cross-modal collaborative representation learning and a large-scale RGBT benchmark for crowd counting[C], 2021 New York IEEE 4823-4833

[[6]] CAO J, PANG Y, XIE J, et al. From handcrafted to deep features for pedestrian detection: a survey [J]. IEEE transactions on pattern analysis and machine intelligence, 2021: 4913–4934.

[[7]] Wang L, Xu L, Yang M H. . Pedestrian detection in crowded scenes via scale and occlusion analysis[C], 2016 New York IEEE 1210-1214

[[8]] Zhang H, Yan C, Li X L, et al.. MSAGNet: multi-stream attribute-guided network for occluded pedestrian detection[J]. IEEE signal processing letter, 2022, 29: 2163-2167,

[[9]] Tang Y, Liu M, Li B Q, et al.. OTP-NMS: toward optimal threshold prediction of NMS for crowded pedestrian detection[J]. IEEE transactions on image processing, 2023, 32: 3176-3187,

[[10]] Xue P, Chen H J, Li Y F, et al.. Multi-scale pedestrian detection with global-local attention and multi-scale receptive field context[J]. IET computer vision, 2023, 17: 13-25,

[[11]] Qing Y, Qian Y Y, Wei H Y, et al.. FE-CSP: a fast and efficient pedestrian detector with center and scale prediction[J]. The journal of supercomputing, 2023, 79(4): 4084-4104,

[[12]] Redmon J, Farhadi A. . YOLO9000: better, faster, stronger[C], 2017 New York IEEE 7263-7271

[[13]] Zhou C, Yuan J. . Bi-box regression for pedestrian detection and occlusion estimation[C], 2018 Berlin, Heidelberg Springer 135-151

[[14]] Zhang S, Wen L, Bian X, et al.. . Occlusion-aware R-CNN: detecting pedestrians in a crowd[C], 2018 Berlin, Heidelberg Springer 637-653

[[15]] Wang X, Xiao T, Jiang Y, et al.. . Repulsion loss: detecting pedestrians in a crowd[C], 2018 New York IEEE 7774-7783

[[16]] Liu S, Huang D, Wang Y. . Adaptive NMS: refining pedestrian detection in a crowd[C], 2019 New York IEEE 6459-6468

[[17]] Woo S, Park J, Lee J Y, et al.. . CBAM: convolutional block attention module[C], 2018 Berlin, Heidelberg Springer 3-19

[[18]] Yao C, Yang J, Ceylan D, et al.. . Learning visibility for robust dense human body estimation[C], 2022 Berlin, Heidelberg Springer 412-428

[[19]] Cheng B, Xiao B, Wang J, et al.. . Higher HRNet: scale-aware representation learning for bottom-up human pose estimation[C], 2020 New York IEEE 5386-5395

[[20]] Lin T Y, Maire M, Belongie S, et al.. . Microsoft coco: common objects in context[C], 2014 Berlin, Heidelberg Springer 740-755

[[21]] Zhang S, Benenson R, Schiele B. . Citypersons: a diverse dataset for pedestrian detection[C], 2017 New York IEEE 3213-3221

[[22]] Cordts M, Omran M, Ramos S, et al.. . The city-scapes dataset for semantic urban scene understanding[C], 2016 New York IEEE 3213-3223

[[23]] Dollár P, Wojek C, Schiele B, et al.. . Pedestrian detection: a benchmark[C], 2009 New York IEEE 304-311

[[24]] Song T, Sun L, Xie D, et al.. . Small-scale pedestrian detection based on somatic topology localization and temporal feature aggregation[C], 2018 Berlin, Heidelberg Springer 1-16

[[25]] Zhang L, Lin L, Liang X, et al.. . Is faster R-CNN doing well for pedestrian detection[C], 2016 Berlin, Heidelberg Springer 443-457

[[26]] Xie J, Pang Y, Khan M H, et al.. Mask-guided attention network and occlusion-sensitive hard example mining for occluded pedestrian detection[J]. IEEE transactions on image processing, 2020, 30: 3872-3884,

[[27]] Liu W, Liao S, Ren W, et al.. . High-level semantic feature detection: a new perspective for pedestrian detection[C], 2019 New York IEEE 5187-5196

[[28]] Zhou P, Zhou C, Peng P, et al.. . NOH-NMS: improving pedestrian detection by nearby objects hallucination[C], 2020 New York ACM 1967-1975

[[29]] M, Zhu C, Wang J, et al.. Adaptive pattern-parameter matching for robust pedestrian detection[C]//Proceedings of the AAAI Conference on Artificial Intelligence, February 2–9, 2021, online. Maynard, Massachusetts: AAAI, 2021, 35(3): 2154-2162

[[30]] Zhang Y, He H, Li J, et al.. . Variational pedestrian detection[C], 2021 New York IEEE 11622-11631

[[31]] Wu J, Zhou C, Yang M, et al.. . Temporal-context enhanced detection of heavily occluded pedestrians[C], 2020 New York IEEE 13430-13439

[[32]] Zou F M, Li X, Xu Q M, et al.. Correlation-and-correction fusion attention network for occluded pedestrian detection[J]. IEEE sensors journal, 2023, 23(6): 6061-6073,