An Integrated Deep Learning Framework for Occluded Pedestrian Tracking

Chen, Kai; Song, Xiao; Zhai, Xiang; Zhang, Baochang; Hou, Baocun; Wang, Yi

doi:10.1109/access.2019.2900296

Cited by 18 publications

(8 citation statements)

References 41 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…It is a task aiming to continually detect a target in a video sequence only its initial position is given. It has numerous realworld applications, including vehicle tracking [4], automatic surveillance [5], and pedestrian tracking [6], [7]. However, it is suffering from some challenging visual attributes, such as background clutters, occlusions [8], motion changes, and size changes.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Multi-Channel Feature Dimension Adaption for Correlation Tracking

Zhang

et al. 2021

IEEE Access

View full text Add to dashboard Cite

Recent discriminative trackers especially based on Correlation Filters (CFs) have shown dominant performance for visual tracking. This kind of trackers benefit from multi-resolution deep features a lot, taking the expressive power of deep Convolutional Neural Networks (CNN). However, distractors in complex scenarios, such as similar targets, occlusion, and deformation, lead to model drift. Meanwhile, learning deep features results in feature redundancy that the increasing number of learning parameters introduces the risk of over-fitting. In this paper, we propose a discriminative CFs based visual tracking method, called dimension adaption correlation filters (DACF). First, the framework adopts the multichannel deep CNN features to obtain a discriminative sample appearance model, resisting the background clutters. Moreover, a dimension adaption operation is introduced to reduce relatively irrelevant parameters as possible, which tackles the issue of over-fitting and promotes the module effectively adapting to different tracking scenes. Furthermore, the DACF formulation optimization can be efficiently performed on the basis of implementing the alternating direction method of multipliers (ADMM). Extensive evaluations are conducted on benchmarks, including OTB2013, OTB2015, VOT2016, and UAV123. The experiments results show that our tracker gains remarkable performance. Especially, DACF obtains an AUC score of 0.698 on OTB2015.INDEX TERMS Correlation filters, multi-channel feature learning, object tracking.

show abstract

Section: Introductionmentioning

confidence: 99%

“…Overall, discriminative trackers are more accurate than generative trackers. Recently, Discriminative Correlation Filters (DCFs) based tracking algorithms have achieved more advancing performance [9]- [11] than traditional discriminative approaches. Two advantages contribute to the advancement of DCF approaches.…”

Section: Introductionmentioning

confidence: 99%

Multi-Channel Feature Dimension Adaption for Correlation Tracking

Zhang

et al. 2021

IEEE Access

View full text Add to dashboard Cite

show abstract

“…This results in frequent mutual occlusions which impede the full-body detection, deteriorating the performance of today's most common detection-based tracking algorithms [1]. Consequently, tracking algorithms have to rely exclusively on the facial region -which is most of the time the only visible part of the subjects-instead of full-body or upper-body regions as in classic pedestrian tracking works [2,3].…”

Section: Introductionmentioning

confidence: 99%

Rank-Based Verification for Long-Term Face Tracking in Crowded Scenes

Barquero¹,

Hupont²,

Fernández³

2021

IEEE Trans. Biom. Behav. Identity Sci.

View full text Add to dashboard Cite

Most current multi-object trackers focus on short-term tracking, and are based on deep and complex systems that often cannot operate in real-time, making them impractical for video-surveillance. In this paper we present a long-term, multi-face tracking architecture conceived for working in crowded contexts where faces are often the only visible part of a person. Our system benefits from advances in the fields of face detection and face recognition to achieve long-term tracking, and is particularly unconstrained to the motion and occlusions of people. It follows a tracking-by-detection approach, combining a fast short-term visual tracker with a novel online tracklet reconnection strategy grounded on rank-based face verification. The proposed rank-based constraint favours higher inter-class distance among tracklets, and reduces the propagation of errors due to wrong reconnections. Additionally, a correction module is included to correct past assignments with no extra computational cost. We present a series of experiments introducing novel specialized metrics for the evaluation of long-term tracking capabilities, and publicly release a video dataset with 10 manually annotated videos and a total length of 8' 54". Our findings validate the robustness of each of the proposed modules, and demonstrate that, in these challenging contexts, our approach yields up to 50% longer tracks than state-of-the-art deep learning trackers.

show abstract

“…Furthermore, the amount of data collected by sensors has exponentially increased with the development of fault detection systems for modern machinery [5], [6]. With its extensive use in target detection, semantic segmentation, autonomous driving [7]- [9], and other fields, deep learning has become the most widely used tool for processing massive amounts of data. The generalized deep learning method employs rich internal information to acquire depth features [10], [11].…”

Section: Introductionmentioning

confidence: 99%

A Novel Transfer Learning Method for Fault Diagnosis Using Maximum Classifier Discrepancy With Marginal Probability Distribution Adaptation

et al. 2020

View full text Add to dashboard Cite

Effective fault diagnosis is essential to ensure the safe and reliable operation of equipment. In recent years, several transfer learning-based methods for diagnosing faults under variable working conditions have been developed. However, these models are designed to completely match the feature distributions between different domains, which is difficult to accomplish because each domain has unique characteristics. To solve this problem, we propose a framework based on the maximum classifier discrepancy with marginal probability distribution adaptation that focuses on task-specific decision boundaries. Specifically, this method captures ambiguous target samples through the predicted discrepancy between two classifiers for the target samples. Furthermore, marginal probability distribution adaptation facilitates the capture of target samples located far from the source domain, and these target samples are brought closer to the source domain through adversarial training. Experimental results indicate that the proposed method demonstrates higher performance and generalization ability than existing fault diagnosis methods.

show abstract

An Integrated Deep Learning Framework for Occluded Pedestrian Tracking

Cited by 18 publications

References 41 publications

Multi-Channel Feature Dimension Adaption for Correlation Tracking

Multi-Channel Feature Dimension Adaption for Correlation Tracking

Rank-Based Verification for Long-Term Face Tracking in Crowded Scenes

A Novel Transfer Learning Method for Fault Diagnosis Using Maximum Classifier Discrepancy With Marginal Probability Distribution Adaptation

Contact Info

Product

Resources

About