It would be nice to see an AI model that can enhance video content from security/fixed cameras where there is no movement from the camera and large portions of the scene are largely repeated across the frames but at varying levels of quality/compression.
