Professor Shao Mingwen from China University of Petroleum (East China) was invited to visit AIDD Center for academic exchanges
In order to encourage teachers and students of our school to take greater steps in scientific research and innovation, promote the cross-integration of different disciplines, and deepen academic exchanges between the mainland and Macau. On the morning of December 2, 2024, the AIDD Center of our college invited Professor Shao Mingwen of China University of Petroleum (East China) to conduct academic exchanges. All teachers and students of the AIDD Center attended the meeting.
Dr. Shao Mingwen is a professor and doctoral supervisor at the School of Computer Science and Technology of China University of Petroleum (East China). He is a member of the Knowledge Engineering and Distributed Intelligence Committee of the Chinese Artificial Intelligence Society, a member of the Artificial Intelligence and Pattern Recognition Committee of the Chinese Computer Society, and a member of the Computer Vision Committee of the Chinese Computer Society. Professor Shao has made extensive achievements in deep learning, computer vision and other fields. In recent years, Professor Shao has published more than 150 academic papers in international journals and academic conferences such as IEEE Trans on Image Processing, AAAI, and ECCV, including more than 120 papers included in SCI and 4 SCI highly cited papers. Professor Shao has successively presided over 5 National Natural Science Foundation projects, 1 project of the Ministry of Education, and 6 provincial and ministerial projects. He serves as the deputy editor-in-chief of the international journal "Journal of Intelligent & Fuzzy Systems" and was selected into the global "Lifetime Scientific Impact" and "Top 2% Top Scientists List". Professor Shao took the actual scientific research experience of the research group as the main line, and took "Research on Image Enhancement Methods in Multi-Degraded Scenarios" as the title, and introduced the basic structure and mathematical principles of the Diffusion model in a simple and easy-to-understand manner, and introduced in detail the application of the Diffusion model in image enhancement in multi-degraded scenarios. Professor Shao first started from the direction of conditional guidance, using the prior knowledge of image color consistency as the condition of the model to guide the Diffusion model to generate enhanced images, solving the problem of image enhancement in low-light scenes. On this basis, Professor Shao's team analyzed the color consistency, exposure distribution and brightness distribution prior knowledge of images taken under different harsh environments, and achieved the restoration and enhancement of images taken under complex and harsh conditions through the sampling method of prior perception combined with the conditional generation guided by transmission. Then, Professor Shao introduced his team's work in the direction of zero-sample image super-resolution. By analyzing the differences between images with different resolutions in the time domain and frequency domain, Professor Shao's team innovatively combined Fourier transform, wavelet transform and cross-attention mechanism. The model fully perceives the image from both the time domain and the frequency domain, effectively overcoming the problem of chessboard artifacts generated when the traditional method enhances the resolution. Finally, Professor Shao's team expanded the research direction to the field of higher-dimensional image processing. Based on Gaussian Splatting, the original Gaussian was decoupled into a pure Gaussian kernel degraded Gaussian through appearance decoupling, and degradation modeling was performed under depth guidance to complete the restoration and enhancement of 3D scenes. In the field of video processing, Professor Shao used a combination of Mask and Diffusion models to perform sample-free flexible and consistent appearance migration on specific areas of the video.
Professor Shao's wonderful report attracted widespread attention and heated discussions from teachers and students present. Many AIDD center doctoral students actively asked questions and exchanged and learned from Professor Shao. Professor Shao answered the students' questions one by one, from the simple to the complex, and gave teachers and students a lot of inspiration. This academic exchange enriched the academic atmosphere of the school, provided teachers and students with valuable exchange and learning opportunities, promoted the intersection and integration of computer and biomedical disciplines, and further deepened the academic exchanges between the mainland and Macau.