Abstract: Deep models, e.g., CNNs and Vision Transformers, have achieved impressive achievements in many vision tasks in the closed world. However, novel classes emerge from time to time in our ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results