Jiahao Wu
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Jiahao Wu has a total of 7 co-authors. Here are the top co-authors by number of shared publications:
- Bo Ma
Beijing Institute of Technology
2 shared publications - Jianbing Shen
University of Macau
1 shared publication - Lianghua Huang
Wan Team, Alibaba Group, Beijing, China
1 shared publication Longyao Liu
1 shared publication- Xin Yi
School of Computer Science, Beijing Institute of Technology, Beijing, China
1 shared publication - Yulin Zhang
School of Computer Science, Beijing Institute of Technology, Beijing, China
1 shared publication
Latest Publications
- Open Access
Two-Step Image Dehazing with Intra-domain and Inter-domain Adaptation
Caused by the difference of data distributions, intra-domain gap and inter-domain gap are widely present in image processing tasks. In the field of image dehazing, certain previous works have paid attention to the inter-domain gap between the synthetic domain and the real domain. However, those methods only establish the connection from the source domain to the target domain without taking into account the large distribution shift within the target domain (intra-domain gap). In this work, we propose a Two-Step Dehazing Network (TSDN) with an intra-domain adaptation and a constrained inter-domain adaptation. First, we subdivide the distributions within the synthetic domain into subsets and mine the optimal subset (easy samples) by loss-based supervision. To alleviate the intra-domain gap of the synthetic domain, we propose an intra-domain adaptation to align distributions of other subsets to the optimal subset by adversarial learning. Finally, we conduct the constrained inter-domain adaptation from the real domain to the optimal subset of the synthetic domain, alleviating the domain shift between domains as well as the distribution shift within the real domain. Extensive experimental results demonstrate that our framework performs favorably against the state-of-the-art algorithms both on the synthetic datasets and the real datasets.
Open AccessTwo-Step Image Dehazing with Intra-domain and Inter-domain Adaptation
Caused by the difference of data distributions, intra-domain gap and inter-domain gap are widely present in image processing tasks. In the field of image dehazing, certain previous works have paid attention to the inter-domain gap between the synthetic domain and the real domain. However, those methods only establish the connection from the source domain to the target domain without taking into account the large distribution shift within the target domain (intra-domain gap). In this work, we propose a Two-Step Dehazing Network (TSDN) with an intra-domain adaptation and a constrained inter-domain adaptation. First, we subdivide the distributions within the synthetic domain into subsets and mine the optimal subset (easy samples) by loss-based supervision. To alleviate the intra-domain gap of the synthetic domain, we propose an intra-domain adaptation to align distributions of other subsets to the optimal subset by adversarial learning. Finally, we conduct the constrained inter-domain adaptation from the real domain to the optimal subset of the synthetic domain, alleviating the domain shift between domains as well as the distribution shift within the real domain. Extensive experimental results demonstrate that our framework performs favorably against the state-of-the-art algorithms both on the synthetic datasets and the real datasets.
- 17 November 2020
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Multimedia
- Vol. 23 (15209210) , 4232-4244
Capturing Relevant Context for Visual Tracking
Studies have shown that context information will promote the robustness of trackers. However, the trackers based on Convolutional Neural Networks (CNNs) only capture local features, hence limiting its performance. Capturing long-range dependencies is crucial for the extraction of much richer features in CNNs, and it has been remarkably successful in numerous computer vision tasks (e.g., object segmentation, pose estimation). In this work, we introduce long-range dependencies into tracking model to capture much richer features, and show the significant improvement in tracking performance. We propose a novel Relevant Context Block (RCB), which employs Graph Convolutional Networks to capture the relevant context. In particular, it selects the k biggest contributors as nodes for each query position (unit) that contain meaningful and discriminative context information, and then updates the nodes by aggregating the differences between query position and its contributors. We find that this operation can be easily incorporated into the existing networks, and can be easily trained using standard back-propagation algorithm in an end-to-end manner. We further develop two trackers, SiamFC and GlobalTrack, to verify the effectiveness of the proposed RCB. Extensive experiments on OTB, VOT, UAV123, LaSOT, TrackingNet, OxUvA, and VOT2018LT show the superiority of our method. For example, Our Siam-RCB outperforms the baseline, SiamFC, by a very large margin (up to 11.2% in success score and 7.8% in precision score) on OTB-100 benchmark.
- 17 November 2020
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Multimedia
- Vol. 23 (15209210) , 4232-4244
Capturing Relevant Context for Visual Tracking
Studies have shown that context information will promote the robustness of trackers. However, the trackers based on Convolutional Neural Networks (CNNs) only capture local features, hence limiting its performance. Capturing long-range dependencies is crucial for the extraction of much richer features in CNNs, and it has been remarkably successful in numerous computer vision tasks (e.g., object segmentation, pose estimation). In this work, we introduce long-range dependencies into tracking model to capture much richer features, and show the significant improvement in tracking performance. We propose a novel Relevant Context Block (RCB), which employs Graph Convolutional Networks to capture the relevant context. In particular, it selects the k biggest contributors as nodes for each query position (unit) that contain meaningful and discriminative context information, and then updates the nodes by aggregating the differences between query position and its contributors. We find that this operation can be easily incorporated into the existing networks, and can be easily trained using standard back-propagation algorithm in an end-to-end manner. We further develop two trackers, SiamFC and GlobalTrack, to verify the effectiveness of the proposed RCB. Extensive experiments on OTB, VOT, UAV123, LaSOT, TrackingNet, OxUvA, and VOT2018LT show the superiority of our method. For example, Our Siam-RCB outperforms the baseline, SiamFC, by a very large margin (up to 11.2% in success score and 7.8% in precision score) on OTB-100 benchmark.