Research Excellence Award

Jisheng Dang is a Chinese computer scientist and academic researcher affiliated with the School of Information Science and Engineering at Lanzhou University. His research primarily focuses on multimodal learning, video understanding, computer vision, video object segmentation, and embodied intelligence. He has contributed to several peer-reviewed publications in internationally recognized journals and conferences, including IEEE Transactions on Image Processing, IEEE Transactions on Neural Networks and Learning Systems, IJCAI, AAAI, and ICME.[1] His scholarly activities additionally include professional reviewing services for major international conferences and journals such as IEEE TPAMI, CVPR, ICML, NeurIPS, ACM MM, and ICLR.[2]

Abstract

Jisheng Dang is a faculty member at Lanzhou University specializing in computer vision, multimodal large language models, and video understanding. His research focuses on video object segmentation, adaptive memory networks, and intelligent visual reasoning systems. He has published scholarly work in internationally recognized journals and conferences, contributing to advancements in multimodal artificial intelligence and efficient video analysis technologies.[3] The article further evaluates his academic contributions and professional suitability for recognition within the framework of the Top Teachers Awards initiative.

Keywords

Computer Vision, Multimodal Learning, Video Understanding, Video Object Segmentation, Artificial Intelligence, Long-Video Understanding, Adaptive Memory Networks, Large Language Models, Intelligent Transportation Systems, Video Data Processing, IEEE Transactions on Image Processing, Neural Networks.

Introduction

The rapid advancement of artificial intelligence and multimodal machine learning has increased the demand for efficient video understanding systems. In this field, Jisheng Dang has contributed to research on long-video processing, adaptive memory modeling, and unified video segmentation frameworks, supporting developments in computer vision and intelligent multimedia analysis.[4]

Jisheng Dang completed his doctoral research at Sun Yat-sen University under Professor Jianhuang Lai and Professor Huicheng Zheng, and later served as a Research Fellow at the NExT++ Laboratory, National University of Singapore, under Professor Tat-Seng Chua. These experiences strengthened his international collaborations in video analysis and multimodal intelligence research.[2]

Research Profile

Jisheng Dang is a tenured Associate Professor at the School of Information Science and Engineering, Lanzhou University, China. His research focuses on multimodal learning, video understanding, video object segmentation, embodied intelligence, and multimodal large language models, with additional contributions in adaptive memory networks and spatiotemporal information processing.[1]

Jisheng Dang actively serves as a reviewer for leading journals and conferences, including IEEE TPAMI, CVPR, ICML, NeurIPS, AAAI, and IJCAI. He has also collaborated with prominent institutions such as the National University of Singapore, Tsinghua University, and Peking University, along with industry partners including Tencent and Huawei.[2]

• Research Areas: Computer Vision, Multimodal Large Language Models, Video Object Segmentation
• Professional Reviewing Experience Across International AI Conferences and Journals

Research Contributions

Jisheng Dang has contributed to efficient and scalable frameworks for video segmentation and multimodal reasoning. His research on adaptive memory systems and spatio-temporal propagation methods improves computational efficiency in long-video processing and video analysis.[3]

His scholarly contributions include the proposal of innovative frameworks such as TW-GRPO, DeSa2VA, and MUPA, designed to improve segmentation accuracy and contextual reasoning in multimodal systems. These approaches attempt to bridge theoretical machine learning research with practical industrial applications involving intelligent transportation systems and advanced multimedia analysis.[5]

• Research on unified video segmentation frameworks for accurate and efficient video object tracking.
• Development of quality-guided dynamic memory approaches for long-video understanding systems.
• Investigation of hallucination mitigation techniques in large video-language models.
• Contributions to multimodal reasoning and adaptive contextual memory architectures.
• Participation in interdisciplinary collaborations linking AI theory with industrial applications.

Publications

Selected publications associated with Jisheng Dang include journal articles and conference papers published in IEEE Transactions on Image Processing, IEEE ICME proceedings, and Neural Networks. Several publications focus on efficient video segmentation, dynamic memory networks, and multimodal understanding systems.[3]

1. Dang, J., Zheng, H., Guo, Y., Lai, J., Hu, B., & Chua, T.-S. (2026). Video Decoupling Networks for Accurate, Efficient, Generalizable, and Robust Video Object Segmentation. IEEE Transactions on Image Processing, Volume 35.
2. Dang, J., Zheng, H., Chen, Z., Li, Z., Guo, Y., & Chua, T.-S. (2026). Fast Track Anything With Sparse Spatio-Temporal Propagation for Unified Video Segmentation. IEEE Transactions on Image Processing, Volume 35.
3. Wang, B., Wen, F., Dang, J., He, H., Wang, X., Zhu, N., & Weng, J. (2025). Mitigating Hallucination in Large Video-Language Models with Injected Semantics. Proceedings of the 2025 IEEE International Conference on Multimedia and Expo (ICME).
4. Wang, B., Jiao, J., Dang, J., Jiang, Q., Lin, J., Chen, Z., Wang, T., & Yang, J. (2025). Quality-Guided Dynamic Memory for LLMs-based Long-Term Video Understanding. Proceedings of the 2025 IEEE International Conference on Multimedia and Expo (ICME).
5. Zhang, L., Dang, J., Zhang, S., Gan, W., Wang, J., Hu, B., Feng, G., & Peng, H. (2026). Graph-enhanced dual low-rank correlation embedding for spatio-temporal EEG fusion in depression recognition. Neural Networks, Volume 198, Article 108609.

Research Impact

The research impact associated with Jisheng Dang is reflected through peer-reviewed publications, scholarly citations, interdisciplinary collaborations, and professional service activities. His work has contributed to research discussions surrounding multimodal large language models, long-video understanding, and scalable segmentation systems.[1]

His publications in IEEE Transactions on Image Processing and conference proceedings have contributed to ongoing advancements in efficient visual processing architectures and multimodal reasoning systems. The application relevance of his research additionally extends to intelligent transportation systems, automated visual understanding, and multimedia analytics.[4]

• Peer-reviewed publications in internationally indexed journals and conferences.
• International collaborations with academic and industrial institutions.
• Reviewer contributions to high-impact AI and computer vision venues.
• Research contributions in video understanding and multimodal AI systems.
• Academic recognition through thesis awards and institutional honors.

Award Suitability

Jisheng Dang has made sustained contributions to computer vision and multimodal machine learning through scholarly publications, collaborative research, and academic service. His publication record and participation in leading international conferences and journals reflect active engagement in the global artificial intelligence research community.[2]

His research in video segmentation, adaptive memory networks, and multimodal understanding systems reflects strong contributions to research excellence and academic innovation. His scholarly achievements and professional collaborations support his recognition within the Top Teachers Awards program.[5]

Conclusion

Jisheng Dang has established a research profile centered on multimodal learning, computer vision, and video understanding technologies. Through scholarly publication, international collaboration, and professional academic service, he has contributed to advancements in video segmentation frameworks and adaptive memory systems for artificial intelligence applications. His academic activities, publication record, and interdisciplinary research collaborations collectively reflect a sustained engagement with contemporary developments in artificial intelligence and multimedia computing research.[1]

External Links

• Scopus Author Profile
• IEEE Xplore Author Profile

References

1. Elsevier. (n.d.). Scopus author details: Jisheng Dang, Author ID 57216844335. Scopus. https://www.scopus.com/authid/detail.uri?authorId=57216844335
2. IEEE. (n.d.). IEEE Xplore Author Profile: Jisheng Dang. IEEE Xplore Digital Library. https://ieeexplore.ieee.org/author/37088932779
3. Dang, J., Zheng, H., Guo, Y., Lai, J., Hu, B., & Chua, T.-S. (2026). Video Decoupling Networks for Accurate, Efficient, Generalizable, and Robust Video Object Segmentation. IEEE Transactions on Image Processing, Volume 35. https://doi.org/10.1109/TIP.2025.3649360
4. Dang, J., Zheng, H., Chen, Z., Li, Z., Guo, Y., & Chua, T.-S. (2026). Fast Track Anything With Sparse Spatio-Temporal Propagation for Unified Video Segmentation. IEEE Transactions on Image Processing, Volume 35. https://doi.org/10.1109/TIP.2025.3649365
5. Zhang, L., Dang, J., Zhang, S., Gan, W., Wang, J., Hu, B., Feng, G., & Peng, H. (2026). Graph-enhanced dual low-rank correlation embedding for spatio-temporal EEG fusion in depression recognition. Neural Networks, Volume 198, Article 108609. https://doi.org/10.1016/j.neunet.2026.108609