Visual sensory anomaly detection (AD) is an essential problem in computer vision, which is gaining momentum recently thanks to the development of AI for good. Compared with semantic anomaly detection which detects anomaly at the label level (semantic shift), visual sensory AD detects the abnormal part of the sample (covariate shift). However, no thorough review has been provided to summarize this area for the computer vision community. In this survey, we are the first one to provide a comprehensive review of visual sensory AD and category into three levels according to the form of anomalies. Furthermore, we classify each kind of anomaly according to the level of supervision. Finally, we summarize the challenges and provide open directions for this community. All resources are available at https://github.com/M-3LAB/awesome-visual-sensory-anomaly-detection.
翻译:视觉感官异常检测(AD)是计算机视觉中的一个基本问题,最近由于发展了AI而正在形成势头。与检测标签层面异常(语义转换)的语义异常异常检测相比,视觉感官AD检测样本的异常部分(交替转换),然而,没有提供彻底的审查来为计算机视觉界总结这方面的情况。在本次调查中,我们是根据异常形式对视觉感官异常和类别进行三级全面审查的第一位。此外,我们根据监督程度对每一种异常进行分类。最后,我们总结了挑战,并为这一群体提供了开放的方向。所有资源都可在https://github.com/M-3LAB/aweome-visual-sensory-anomary-detraction上查阅。