Situation-adaptive neural network for fast pre-computing image enhancement

作     者：Xinyue LI Huiyu DUAN Jia WANG Xiaohong LIU Yitong CHEN Guangtao ZHAI 

作者机构：Institute of Image Communication and Network Engineering Department of Electronic EngineeringShanghai Jiao Tong University 

出 版 物：《Science China(Information Sciences)》 (中国科学:信息科学(英文版))

年 卷 期：2025年

核心收录：

学科分类：12[管理学] 1201[管理学-管理科学与工程(可授管理学、工学学位)] 081104[工学-模式识别与智能系统] 08[工学] 080203[工学-机械设计及理论] 0835[工学-软件工程] 0802[工学-机械工程] 0811[工学-控制科学与工程] 0812[工学-计算机科学与技术（可授工学、理学学位）] 

基　　金：supported in part by National Natural Science Foundation of China (Grant No. 62301310) Shanghai Pujiang Program (Grant No. 22PJ1406800) 

摘      要：As intelligent vision tasks become more widespread, enhancing image quality before further computational analysis is crucial. Recently, deep learning has shown potential for automated pre-computing enhancement, but it typically requires substantial computational resources and is hard to adapt to in multiple situations without *** practice, image enhancement often demands flexible adjustments based on different situations and subsequent computation devices, such as optical computing. Therefore,we propose SAEnhancer, a situation-adaptive neural network for fast pre-computing image enhancement. It learns from a small sample set to achieve personalized and adaptive enhancements for various situations without re-training,using semantic-aware embedding for precise color adjustments, surpassing traditional 3D lookup tables (LUTs), and enhancing computational effectiveness in intelligent vision applications.