Yin Cui
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Yin Cui has a total of 9 co-authors. Here are the top co-authors by number of shared publications:
- Hongge Ren
School of Control and Mechanical Engineering, Tianjin Chengjian University, Tianjin 300384, China
2 shared publications - Boqing Gong
Boston University
1 shared publication - Florian Schroff
Google research
1 shared publication - Hartwig Adam
Google research
1 shared publication - Liangzhe Yuan
Google research
1 shared publication - Ming-Hsuan Yang
University of California, Merced, USA
1 shared publication
Latest Publications
- Open Access
- 1 June 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 13957-13966
Contextualized Spatio-Temporal Contrastive Learning with Self-Supervision
Modern self-supervised learning algorithms typically enforce persistency of instance representations across views. While being very effective on learning holistic image and video representations, such an objective becomes suboptimal for learning spatio-temporally fine-grained features in videos, where scenes and instances evolve through space and time. In this paper, we present Contextualized Spatio-Temporal Contrastive Learning (ConST-CL) to effectively learn spatio-temporally fine-grained video representations via self-supervision. We first design a region-based pretext task which requires the model to transform instance representations from one view to another, guided by context features. Further, we introduce a simple network design that successfully reconciles the simultaneous learning process of both holistic and local representations. We evaluate our learned representations on a variety of downstream tasks and show that ConST-CL achieves competitive results on 6 datasets, including Kinetics, UCF, HMDB, AVA-Kinetics, AVA and OTB. Our code and models will be available at https://github.com/tensorflow/models/tree/master/official/projects/const_cl.
- 1 June 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 13957-13966
Open AccessContextualized Spatio-Temporal Contrastive Learning with Self-Supervision
Modern self-supervised learning algorithms typically enforce persistency of instance representations across views. While being very effective on learning holistic image and video representations, such an objective becomes suboptimal for learning spatio-temporally fine-grained features in videos, where scenes and instances evolve through space and time. In this paper, we present Contextualized Spatio-Temporal Contrastive Learning (ConST-CL) to effectively learn spatio-temporally fine-grained video representations via self-supervision. We first design a region-based pretext task which requires the model to transform instance representations from one view to another, guided by context features. Further, we introduce a simple network design that successfully reconciles the simultaneous learning process of both holistic and local representations. We evaluate our learned representations on a variety of downstream tasks and show that ConST-CL achieves competitive results on 6 datasets, including Kinetics, UCF, HMDB, AVA-Kinetics, AVA and OTB. Our code and models will be available at https://github.com/tensorflow/models/tree/master/official/projects/const_cl.
- Open Access
- 23 December 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 10, 2457-2468
Research on Anomaly Suppression Correlation Filtering Algorithm
Aiming at the problems of target occlusion, illumination change, and fast motion in target tracking, an anomaly suppression correlation filtering algorithm is proposed. The histogram of oriented gradient (HOG) feature and color name (CN) feature are extracted to construct the target appearance model, and an abnormal suppression correlation filter is constructed to suppress sudden changes in the filter response graph. The alternating direction method of multipliers (ADMM) method is used to speed up the calculation of the filter. A scale filter is added in the article to solve the problem of inaccurate tracking caused by a single scale. In the model update process, the three-layer rotating circle memory model is used to update, which improves the tracking ability of the algorithm. It is better to update the target model in the scene where the target is occluded. The proposed algorithm and BACF, KCC, KCF, MKCFup, ARCF are tested in the OTB50, OTB100, UAV123, TC128 experimental data sets. The results show that the visual tracking algorithm proposed in the article has a high success rate and accuracy, and has certain research value.
- 23 December 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 10, 2457-2468
Open AccessResearch on Anomaly Suppression Correlation Filtering Algorithm
Aiming at the problems of target occlusion, illumination change, and fast motion in target tracking, an anomaly suppression correlation filtering algorithm is proposed. The histogram of oriented gradient (HOG) feature and color name (CN) feature are extracted to construct the target appearance model, and an abnormal suppression correlation filter is constructed to suppress sudden changes in the filter response graph. The alternating direction method of multipliers (ADMM) method is used to speed up the calculation of the filter. A scale filter is added in the article to solve the problem of inaccurate tracking caused by a single scale. In the model update process, the three-layer rotating circle memory model is used to update, which improves the tracking ability of the algorithm. It is better to update the target model in the scene where the target is occluded. The proposed algorithm and BACF, KCC, KCF, MKCFup, ARCF are tested in the OTB50, OTB100, UAV123, TC128 experimental data sets. The results show that the visual tracking algorithm proposed in the article has a high success rate and accuracy, and has certain research value.
- Open Access
- 27 July 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 9, 105318-105326
Research on Visual Tracking Algorithm Based on Peak Sidelobe Ratio
Aiming at the problems of illumination changes, target deformation and background clutter in the target tracking field, a visual tracking algorithm based on peak sidelobe ratio is proposed. The object-interference model is used to represent the target appearance model, and context information is added to the relevant filtering framework. Training the filter internally to enhance the ability of filter discrimination. In the model update process, it is easier to introduce samples that cannot characterize the target, and use peak sidelobe comparison to update the tracking parameters, which can enhance the generalization ability of the model. Tested with some classic and recently algorithms in the OTB50, OTB100, UAV123, TC128 experimental video data set, the experiment’ results show that the visual tracking algorithm that is proposed in the article can track the target more accurately. It has important research on the development of intelligent video surveillance value.
- 27 July 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 9, 105318-105326
Open AccessResearch on Visual Tracking Algorithm Based on Peak Sidelobe Ratio
Aiming at the problems of illumination changes, target deformation and background clutter in the target tracking field, a visual tracking algorithm based on peak sidelobe ratio is proposed. The object-interference model is used to represent the target appearance model, and context information is added to the relevant filtering framework. Training the filter internally to enhance the ability of filter discrimination. In the model update process, it is easier to introduce samples that cannot characterize the target, and use peak sidelobe comparison to update the tracking parameters, which can enhance the generalization ability of the model. Tested with some classic and recently algorithms in the OTB50, OTB100, UAV123, TC128 experimental video data set, the experiment’ results show that the visual tracking algorithm that is proposed in the article can track the target more accurately. It has important research on the development of intelligent video surveillance value.