Open World Object Detection

Open-world object detection is a challenging problem in the field of computer vision, with its core objectives being: 1) to identify unseen objects and label them as "unknown" without explicit supervision; 2) to incrementally learn these unknown categories upon receiving the corresponding labels, while avoiding forgetting the learned classes. The application value of this task lies in enhancing the model's adaptability and robustness in real-world environments, and improving its generalization capability.