“Related research has made new progress in the field of movie scene segmentation. Experts have proposed the HASSNet hybrid architecture method, which significantly improves scene segmentation accuracy through pre training and fine-tuning strategies, combined with state space models and self attention mechanisms, providing strong support for movie content understanding and multimedia applications.”