Worm Video Denoising and Detection

There are significant spatial and temporal variations in the worm, corona ring, and food noise. Simply applying deep learning methods may not yield satisfactory results. Therefore, we aim to first denoise the worm video frame by frame, specifically targeting the removal of corona and food noise using an adaptive compressive sensing algorithm. This approach will allow us to recover a cleaner image of the worm, which can then be fed into a deep learning model for detection.