Balancing Shared and Task-Specific Representations: A Hybrid Approach to Depth-Aware Video Panoptic Segmentation Paper • 2412.07966 • Published Dec 10, 2024 • 1