% Describe-to-Detect (D2D) leverages successful descriptor models without the need for any additional training. Our method selects keypoints as ...
確定! 回上一頁