Yu Zhang

Micro-video Scene Classification Method Based on Generative Feature Disentanglement and Multimodal Dynamical Fusion

Since micro-video is typically captured in complex real-world environments, it often contains a large amount of noise and redundant information. They can weaken the expression of key semantics, thereby affecting the model's accuracy in scene recognition and classification. Owing to the production, editing, or other inf...