Unlike the existing work such as soft-quantization, which only focuses on the Euclidean (l2) distance in descriptor space, we utilize the visual ...
確定! 回上一頁