Novel techniques extract more accurate data from images degraded by environmental factors

A team of researchers has developed novel approaches using computer vision and deep learning to resolve the problem of low-level vision in videos caused by rain and night-time conditions, as well as improve the accuracy of 3D human pose estimation in videos.

source https://www.sciencedaily.com/releases/2021/07/210719110548.htm