Density Map Regression Guided Detection Network for RGB-D Crowd Counting and Localization

Lian, Dongze; Li, Jing; Zheng, Jianming; Luo, Weixin; Gao, Shenghua

doi:10.1109/cvpr.2019.00192

Cited by 160 publications

(117 citation statements)

References 24 publications

Supporting

Mentioning

117

Contrasting

Order By: Relevance

“…(b) Low quality depth images are cluttered thus may be harmful for the prediction. also proven to be an effective way in the applications of object detection [27], semantic segmentation [32], and crowd counting [31].…”

Section: Introductionmentioning

confidence: 99%

Progressively Complementarity-Aware Fusion Network for RGB-D Salient Object Detection

Chen

2018

2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition

303

315

View full text Add to dashboard Cite

In this paper, we aim to develop an efficient and compact deep network for RGB-D salient object detection, where the depth image provides complementary information to boost performance in complex scenarios. Starting from a coarse initial prediction by a multi-scale residual block, we propose a progressively guided alternate refinement network to refine it. Instead of using ImageNet pre-trained backbone network, we first construct a lightweight depth stream by learning from scratch, which can extract complementary features more efficiently with less redundancy. Then, different from the existing fusion based methods, RGB and depth features are fed into proposed guided residual (GR) blocks alternately to reduce their mutual degradation. By assigning progressive guidance in the stacked GR blocks within each side-output, the false detection and missing parts can be well remedied. Extensive experiments on seven benchmark datasets demonstrate that our model outperforms existing state-of-the-art approaches by a large margin, and also shows superiority in efficiency (71 FPS) and model size (64.9 MB).

show abstract

Section: Introductionmentioning

confidence: 99%

Progressively Complementarity-Aware Fusion Network for RGB-D Salient Object Detection

Chen

2018

2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition

303

315

View full text Add to dashboard Cite

show abstract

“…Only using image lowlevel information, these methods had high efficiencies, but their performance was far from satisfactory for real-world applications. Recently, we have witnessed the great success of convolutional neural networks [29,32,43,48,53,55,59,63,66,71,72] in crowd counting. Most of these previous approaches focused on how to improve the performance of deep models.…”

Section: Related Workmentioning

confidence: 99%

Efficient Crowd Counting via Structured Knowledge Transfer

Liu

Chen

et al. 2020

Proceedings of the 28th ACM International Conference on Multimedia

View full text Add to dashboard Cite

Crowd counting is an application-oriented task and its inference efficiency is crucial for real-world applications. However, most previous works relied on heavy backbone networks and required prohibitive run-time consumption, which would seriously restrict their deployment scopes and cause poor scalability. To liberate these crowd counting models, we propose a novel Structured Knowledge Transfer (SKT) framework, which fully exploits the structured knowledge of a well-trained teacher network to generate a lightweight but still highly effective student network. Specifically, it is integrated with two complementary transfer modules, including an Intra-Layer Pattern Transfer which sequentially distills the knowledge embedded in layer-wise features of the teacher network to guide feature learning of the student network and an Inter-Layer Relation Transfer which densely distills the cross-layer correlation knowledge of the teacher to regularize the student's feature evolution. Consequently, our student network can derive the layer-wise and cross-layer knowledge from the teacher network to learn compact yet effective features. Extensive evaluations on three benchmarks well demonstrate the effectiveness of our SKT for extensive crowd counting models. In particular, only using around 6% of the parameters and computation cost of original models, our distilled VGG-based models obtain at least 6.5× speed-up on an Nvidia 1080 GPU and even achieve state-of-the-art performance. Our code and models are available at https://github.com/HCPLab-SYSU/SKT.

show abstract

“…They were commonly adopted in relatively low-density crowds [11,13] as the performance would decay severely in high-density crowds with small and occluded persons. A recent resurgence of detection-based methods in crowd counting [15,17,22,32] is owing to the advances of object detection in the deep learning context [16,19,24,25]. [17] trained an end-to-end people detector for crowded scenes depending on annotations of bounding boxes of persons.…”

Section: Detection-based Crowd Countingmentioning

confidence: 99%

“…Despite the resurgence of detection-based methods, in terms of counting accuracy in dense crowds, they are still not as competitive as those regression-based methods, and often need to be integrated into the latter [15,17,22]. The integration can be through the attention module in an implicit way [15,17], while in this paper, we distill the bi-directional knowledge amid regression and detectionbased models and learn to explicitly transform the output from one to the other in unsupervised crowd counting.…”

Section: Detection-based Crowd Countingmentioning

confidence: 99%

“…These methods are mostly realized via the density regression-based models, where a density distribution is learnt for each crowd image whose integral over the density map gives the total count of that image. A counterpart direction, which has not been investigated in the cross-domain setting, is the detection-based methods [15,22], where every individual is to be localized in the crowd images. The regression-based methods perform very well in high-density and congested crowd scenes as they do not predict individual locations in particular with occlusions and overlapping.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Towards Unsupervised Crowd Counting via Regression-Detection Bi-knowledge Transfer

Liu

Wang

Shi

et al. 2020

Proceedings of the 28th ACM International Conference on Multimedia

View full text Add to dashboard Cite

Unsupervised crowd counting is a challenging yet not largely explored task. In this paper, we explore it in a transfer learning setting where we learn to detect and count persons in an unlabeled target set by transferring bi-knowledge learnt from regression-and detection-based models in a labeled source set. The dual source knowledge of the two models is heterogeneous and complementary as they capture different modalities of the crowd distribution. We formulate the mutual transformations between the outputs of regression-and detection-based models as two scene-agnostic transformers which enable knowledge distillation between the two models. Given the regression-and detection-based models and their mutual transformers learnt in the source, we introduce an iterative self-supervised learning scheme with regression-detection bi-knowledge transfer in the target. Extensive experiments on standard crowd counting benchmarks, ShanghaiTech, UCF_CC_50, and UCF_QNRF demonstrate a substantial improvement of our method over other state-of-the-arts in the transfer learning setting. CCS CONCEPTS • Information systems → Multimedia information systems; • Humancentered computing → Collaborative and social computing; Visualization; • Computing methodologies → Computer vision.

show abstract

Density Map Regression Guided Detection Network for RGB-D Crowd Counting and Localization

Cited by 160 publications

References 24 publications

Progressively Complementarity-Aware Fusion Network for RGB-D Salient Object Detection

Progressively Complementarity-Aware Fusion Network for RGB-D Salient Object Detection

Efficient Crowd Counting via Structured Knowledge Transfer

Towards Unsupervised Crowd Counting via Regression-Detection Bi-knowledge Transfer

Contact Info

Product

Resources

About