Occluded Pedestrian Detection Through Guided Attention in CNNs

By Shanshan Zhang, Jian Yang, Bernt Schiele. 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2018

359 citations recorded by the source corpus.

Pedestrian detection has progressed significantly in the last years. However, occluded people are notoriously hard to detect, as their appearance varies substantially depending on a wide range of occlusion patterns. In this paper, we aim to propose a simple and compact method based on the FasterRCNN architecture for occluded pedestrian detection. We start with interpreting CNN channel features of a pedestrian detector, and we find that different channels activate responses for different body parts respectively. These findings motivate us to employ an attention mechanism across channels to represent various occlusion patterns in one single model, as each occlusion pattern can be formulated as some specific combination of body parts. Therefore, an attention network with self or external guidances is proposed as an add-on to the baseline FasterRCNN detector. When evaluating on the heavy occlusion subset, we achieve a significant improvement of 8pp to the baseline FasterRCNN detector on CityPersons and on Caltech we outperform the state-of-the-art method by 4pp.

Subfields: Computer Vision and Pattern Recognition, Safety, Risk, Reliability and Quality

View Paper: https://www.semanticscholar.org/paper/63f97279a463a3bd18cc3c284d7a8ab28a5aaa1d

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

In Safety, Risk, Reliability and Quality

Citations in the corpus

A survey on instance segmentation: state of the art (2020, background): “[61] also proposed training of an adversarial network [62].”
Pose-Guided Feature Alignment for Occluded Person Re-Identification (2019, methodology): “[41] also utilize the pose-guided attention mechanism to tackle the occlusion problems.”
Illumination-aware Faster R-CNN for Robust Multispectral Pedestrian Detection (2018, methodology): “Recent work [32, 33] were built upon Faster R-CNN architecture and focused on occlusion handling.”
Detection in Crowded Scenes: One Proposal, Multiple Predictions (2020, background): “There are other approaches on crowded detection, for example, part-based detectors [44, 38, 46, 3], which is mainly used in detecting special instances such as pedestrian.”
KGSNet: Key-Point-Guided Super-Resolution Network for Pedestrian Detection in the Wild (2020, methodology): “Note that the results of PDOE [38], RepLoss [25] and ALFNet [63] DeepParts [28], F-DNN + SS [20], FasterRCNN + ATT [40], CompACT-Deep [78], SA-FastRCNN [61], RPN + BF [22], MS-CNN [79], GDFL [39], TLL-TFA [41], SDS-RCNN [19], PDOE [38], RepLoss [25], and ALFNet [63].”
Predictive Inequity in Object Detection (2019, background): “Recent works have addressed this is by incorporating multiple representations into their model, one for each pedestrian scale (Li et al., 2018), or for different body parts to handle occlusion (Zhang et al., 2018).”
Improving Multispectral Pedestrian Detection by Addressing Modality Imbalance Problems (2020, methodology): “ile maintaining the fastest speed. 2 Relate Work 2.1 Multispectral Pedestrian Detection CNN-based pedestrian detection has achieved notable progress in recent years with methods of occlusion handling [46,32,38], cascaded detection systems [30,2], semantic attention [49,3], anchor-free approach [31],”
Deep Learning for Generic Object Detection: A Survey (2018, background): “ly ﬁxed geometric structures. Wang et al. [280] propose to learn an adversarial network that generates examples with occlusions and deformations, and context may be helpful in dealing with occlusions [309]. Despite these efforts, the occlusion problem is far from being solved; applying GANs to this ”
Temporal-Context Enhanced Detection of Heavily Occluded Pedestrians (2020, methodology): “To alleviate the occlusion issue, recent methods are designed by exploiting attention mechanism [29, 9], feature transformation [11] and part-based detection [22, 19, 13].”
Pose2Seg: Detection Free Human Instance Segmentation (2018, background): “There has been much good research in the ﬁelds of human pose estimation [1, 2, 6, 14, 20, 26, 40], pedestrian detection [25, 41, 42], portrait segmentation [35, 36, 37], and face recognition [18, 23, 24, 27, 39, 43, 44], much of which has already produced practical value in real life.”
Interaction-And-Aggregation Network for Person Re-Identification (2019, background): “[58] have discovered that most channel maps of highlevel features show strong responses for specific parts.”
Adaptive NMS: Refining Pedestrian Detection in a Crowd (2019, background): “[50] employs an attention mechanism across channels to represent various occlusion patterns.”