SmartEraser is a object removal model built on the novel Masked-Region Guidance paradigm. SmartEraser outperforms existing methods by smartly identifying the target object to remove while effectively ...
Abstract: Low-light image augmentation is crucial in many applications where visibility is regularly impeded by suboptimal lighting, such as autonomous driving, surveillance, and medical imaging.
2025.09.26: Release the real-world EDTR detection model and demo examples. The code has also been simplified. 2025.09.01: The code is released. Our code has been tested with PyTorch 2.2.2 and CUDA ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results