The Research on Green Screen Video Segmentation Based on Multi-scale Deep Learning Techniques

As a key branch of artificial intelligence and machine learning, deep learning technology has continuously developed and improved, leading to significant progress and widespread applications in the field of video segmentation. However, within this technological wave, traditional green screen video segmentation methods seem to have been marginalized, and related research has received relatively little attention. To address the shortcomings of existing research, this approach proposes a new green screen video segmentation method that combines traditional chroma keying techniques with the lightweight deep learning network MobileNetV4, aiming to achieve efficient and accurate segmentation results. The chroma key module generates a rough mask, providing auxiliary information for subsequent feature extraction, while feature fusion and a recurrent decoder module enhance the consistency and detail retention between video frames. Experimental results demonstrate that this method performs excellently on both the VideoMatte240K dataset and a self-constructed dataset, effectively improving the precision and efficiency of video segmentation. The region similarity J reached 98%, and the contour accuracy F reached 94%, surpassing the average value Avg(J&F) of the second-ranked MODNet algorithm by 9.6%.