Multi-label Classification with Panoptic Context Aggregation Networks

Mingyuan Jiu, Hailong Zhu, Wenchuan Wei +3

Context modeling is crucial for visual recognition, enabling highly discriminative image representations by integrating both intrinsic and extrinsic relationships between objects a…