Jing Xu
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
Research Subjects
Jing Xu has a total of 10 co-authors. Here are the top co-authors by number of shared publications:
- Pan Gao
Nanjing University of Aeronautics and Astronautics, Nanjing, China
2 shared publications - Weikai Yang
Data Science and Analytics Thrust, Hong Kong University of Science and Technology (Guangzhou), Guangzhou, Guangdong, China
2 shared publications - Wentao Shi
College of Computer Science and Technology, Nanjing University of Aeronautics and Astronautics, Nanjing, China
2 shared publications Jie Qin
College of Artificial Intelligence, Nanjing University of Aeronautics and Astronautics, Nanjing, China
1 shared publication- Lijuan Zhang
College of Electronic and Information Engineering, Nanjing University of Aeronautics and Astronautics, Nanjing, China
1 shared publication Qizhu Li
TikTok Pte. Ltd., Singapore
1 shared publication
Latest Publications
- Open Access
- 9 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 921-929
RISE: Reliable Inference and State Evolution for CLIP Test-Time Adaptation under Mixed Corruption Streams
While CLIP excels at zero-shot recognition, its accuracy degrades sharply under common corruptions such as noise, blur, weather effects, and digital distortions. To mitigate this issue, existing test-time adaptation (TTA) methods adapt CLIP during inference, but they are largely developed and evaluated under isolated-corruption protocols, where each corruption type is treated as an independent test condition. These methods can fail when handling mixed online test streams, in which corruption types can change frequently across batches and adaptation errors can accumulate over time. Motivated by this limitation, we introduce Hierarchical Mixed Corruption Streams (HMCS), a protocol family that systematically evaluates CLIP-TTA under progressively broader within-family, family-pair, and global mixing scenarios. We further propose Reliable Inference and State Evolution (RISE), a CLIP-TTA framework that decouples batch-local adaptation from persistent non-parametric state while coordinating them through shared online statistics. Built on a batch-local adaptation step, RISE improves long-horizon stability through filtered memory updates, shift-aware stability control, and quality-aware prediction fusion. Extensive experiments on CIFAR-10-C, CIFAR-100-C, and ImageNet-C-5K show that RISE consistently outperforms strong CLIP-TTA baselines, suggesting that under mixed corruption streams with frequent corruption changes, effective adaptation depends not only on the current batch but also on preserving the reliability of accumulated state.
- 9 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 921-929
Open AccessRISE: Reliable Inference and State Evolution for CLIP Test-Time Adaptation under Mixed Corruption Streams
While CLIP excels at zero-shot recognition, its accuracy degrades sharply under common corruptions such as noise, blur, weather effects, and digital distortions. To mitigate this issue, existing test-time adaptation (TTA) methods adapt CLIP during inference, but they are largely developed and evaluated under isolated-corruption protocols, where each corruption type is treated as an independent test condition. These methods can fail when handling mixed online test streams, in which corruption types can change frequently across batches and adaptation errors can accumulate over time. Motivated by this limitation, we introduce Hierarchical Mixed Corruption Streams (HMCS), a protocol family that systematically evaluates CLIP-TTA under progressively broader within-family, family-pair, and global mixing scenarios. We further propose Reliable Inference and State Evolution (RISE), a CLIP-TTA framework that decouples batch-local adaptation from persistent non-parametric state while coordinating them through shared online statistics. Built on a batch-local adaptation step, RISE improves long-horizon stability through filtered memory updates, shift-aware stability control, and quality-aware prediction fusion. Extensive experiments on CIFAR-10-C, CIFAR-100-C, and ImageNet-C-5K show that RISE consistently outperforms strong CLIP-TTA baselines, suggesting that under mixed corruption streams with frequent corruption changes, effective adaptation depends not only on the current batch but also on preserving the reliability of accumulated state.
- 22 December 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Intelligent Transportation Systems
- Vol. 27 (4) , 4428-4437
Fast Wrong-Way Cycling Detection in CCTV Videos: Sparse Sampling Is All You Need
Effective monitoring of unusual transportation behaviors, such as wrong-way cycling (i.e., riding a bicycle or e-bike against designated traffic flow), is crucial for optimizing law enforcement deployment and traffic planning. However, accurately recording all wrong-way cycling events is both unnecessary and infeasible in resource-constrained environments, as it requires high-resolution cameras for evidence collection and event detection. To address this challenge, we propose WWC-Predictor, a novel method for efficiently estimating the wrong-way cycling ratio, defined as the proportion of wrong-way cycling events relative to the total number of cycling movements over a given time period. The core innovation of our method lies in accurately detecting wrong-way cycling events in sparsely sampled frames using a light-weight detector, then estimating the overall ratio using an autoregressive moving average model. To evaluate the effectiveness of our method, we construct a benchmark dataset consisting of 35 minutes of video sequences with minute-level annotations. Our method achieves an average error rate of a mere 1.475% while consuming only 19.12% GPU time required by conventional tracking methods, validating its effectiveness in estimating the wrong-way cycling ratio. Our source code is publicly available at: https://github.com/VICA-Lab-HKUST-GZ/WWC-Predictor
- 22 December 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Intelligent Transportation Systems
- Vol. 27 (4) , 4428-4437
Fast Wrong-Way Cycling Detection in CCTV Videos: Sparse Sampling Is All You Need
Effective monitoring of unusual transportation behaviors, such as wrong-way cycling (i.e., riding a bicycle or e-bike against designated traffic flow), is crucial for optimizing law enforcement deployment and traffic planning. However, accurately recording all wrong-way cycling events is both unnecessary and infeasible in resource-constrained environments, as it requires high-resolution cameras for evidence collection and event detection. To address this challenge, we propose WWC-Predictor, a novel method for efficiently estimating the wrong-way cycling ratio, defined as the proportion of wrong-way cycling events relative to the total number of cycling movements over a given time period. The core innovation of our method lies in accurately detecting wrong-way cycling events in sparsely sampled frames using a light-weight detector, then estimating the overall ratio using an autoregressive moving average model. To evaluate the effectiveness of our method, we construct a benchmark dataset consisting of 35 minutes of video sequences with minute-level annotations. Our method achieves an average error rate of a mere 1.475% while consuming only 19.12% GPU time required by conventional tracking methods, validating its effectiveness in estimating the wrong-way cycling ratio. Our source code is publicly available at: https://github.com/VICA-Lab-HKUST-GZ/WWC-Predictor
- 9 September 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Emerging Topics in Computational Intelligence
- Vol. 9 (1) , 202-212
MUSTER: A Multi-Scale Transformer-Based Decoder for Semantic Segmentation
In recent works on semantic segmentation, there has been a significant focus on designing and integrating transformer-based encoders. However, less attention has been given to transformer-based decoders. We emphasize that the decoder stage is equally vital as the encoder in achieving superior segmentation performance. It disentangles and refines high-level cues, enabling precise object boundary delineation at the pixel level. In this paper, we introduce a novel transformer-based decoder called MUSTER, which seamlessly integrates with hierarchical encoders and consistently delivers high-quality segmentation results, regardless of the encoder architecture. Furthermore, we present a variant of MUSTER that reduces FLOPS while maintaining performance. MUSTER incorporates carefully designed multi-head skip attention (MSKA) units and introduces innovative upsampling operations. The MSKA units enable the fusion of multi-scale features from the encoder and decoder, facilitating comprehensive information integration. The upsampling operation leverages encoder features to enhance object localization and surpasses traditional upsampling methods, improving mIoU (mean Intersection over Union) by 0.4% to 3.2%. On the challenging ADE20K dataset, our best model achieves a single-scale mIoU of 50.23 and a multi-scale mIoU of 51.88, which is on-par with the current state-of-the-art model. Remarkably, we achieve this while significantly reducing the number of FLOPs by 61.3%.
- 9 September 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Emerging Topics in Computational Intelligence
- Vol. 9 (1) , 202-212
MUSTER: A Multi-Scale Transformer-Based Decoder for Semantic Segmentation
In recent works on semantic segmentation, there has been a significant focus on designing and integrating transformer-based encoders. However, less attention has been given to transformer-based decoders. We emphasize that the decoder stage is equally vital as the encoder in achieving superior segmentation performance. It disentangles and refines high-level cues, enabling precise object boundary delineation at the pixel level. In this paper, we introduce a novel transformer-based decoder called MUSTER, which seamlessly integrates with hierarchical encoders and consistently delivers high-quality segmentation results, regardless of the encoder architecture. Furthermore, we present a variant of MUSTER that reduces FLOPS while maintaining performance. MUSTER incorporates carefully designed multi-head skip attention (MSKA) units and introduces innovative upsampling operations. The MSKA units enable the fusion of multi-scale features from the encoder and decoder, facilitating comprehensive information integration. The upsampling operation leverages encoder features to enhance object localization and surpasses traditional upsampling methods, improving mIoU (mean Intersection over Union) by 0.4% to 3.2%. On the challenging ADE20K dataset, our best model achieves a single-scale mIoU of 50.23 and a multi-scale mIoU of 51.88, which is on-par with the current state-of-the-art model. Remarkably, we achieve this while significantly reducing the number of FLOPs by 61.3%.