Recently, Professor Sean He, the Deputy Head of CS School and the Director of Computer Vision and Intelligent Perception Laboratory, had his article titled “Pixels, Regions, and Objects: Multiple Enhancement for Salient Object Detection” accepted by CVPR as a corresponding author. This article was co-authored by Yi Wang and Xin Fan of Dalian University of Technology, as well as Ruili Wang of Massey University, New Zealand. In the article, a flexible multi-scale feature enhancement module (ME-Module) is proposed to aggregate and enhance features through atrous spatial pyramid pooling and global-local attention mechanisms. The article solves the problem of incomplete and blurry boundaries of targets in complex scenes caused by most Salient Object Detection (SOD) methods interpreting images from the perspective of pixels and using simple end-to-end training methods to learn salient features.

About 80% of the information that human’s perception in the environment comes from vision. The Multiple Enhancement Network (MENet) used in this article is an innovative network that is based on the iterative refinement mechanism and content integrity of the human visual system. When compared to the previous optimal model, the error can be reduced by 2-10% when using this system for detection, significantly enhancing the integrity and accuracy of the process.

CVPR, i.e., the IEEE International Conference on Computer Vision and Pattern Recognition, is the world's No.1 computer vision conference. The overall acceptance rate for CVPR 2023 is 25.78%. In the 2022 List of Top Publications as per Google Scholar Metric, CVPR ranks fourth, only behind Nature, NEJM, and Science, the three top journals in the world. Among the international journals and conferences recommended by the China Computer Federation, CVPR is also a Class A conference in the field of Artificial Intelligence.

Professor Sean He has been deeply involved in the fields of computer vision, machine learning, data analysis, and pattern recognition, and has achieved a series of important research results. According to Google Scholar, his published papers have been cited 9931 times. He has published papers in top journals in this field, including ESI Highly Cited Papers, and the articles appearing in top-tier journals such as IEEE TPAMI, ACM Computing Surveys and in top academic conferences such as CVPR, AAAI, ACL, ECCV and ACM MM. It is his research spirit of in-depth exploration and practical work that has enabled him to achieve rich research results in his research field, and be included in the “World’s Top 2% Scientists” list released by Stanford University. With the support of the research resources of the UNNC and the advantages of the University of Nottingham, he has led the “Computer Vision and Intelligent Perception Laboratory” (CVIP), which has established a core research team and gathered 11 outstanding young and middle-aged leading academics and 33 PhD students to continuously expand and extend research areas.