Thus, large numbers of unlabeled text images can be adopted to markedly enhance the discrimination and robustness of our font classifier.
確定! 回上一頁