On November 14th,2023, Dai Qionghai, President of Chinese Association for Artificial Intelligence (CAAI), along with Yu Li, and Wu Jiamin from Tsinghua University, jointly published a paper titled “Bio-friendly long-term subcellular dynamic recording by self-supervised image enhancement microscopy”. The research team introduced DeepSemi, an image enhancement framework based on self-supervised learning. By incorporating newly designed blind-spot neural network, the DeepSeMi can effectively enhance microscopic images without sacrificing spatiotemporal resolution. Through comprehensive validation on various samples and instruments, the three researchers demonstrated that the DeepSeMi is a kind of versatile and biocompatible tool which is capable of overcoming limitations of photon noise.