Images were then cropped by align- ing faces according to predefined image composition tem- plates [27]. Ciocca et al. combined visual saliency informa- tion ...
確定! 回上一頁