method: HoText_v02018-01-10
Authors: Zijian Zhou, Zhizhong Su, Shuangping Huang
Description: This result is just use 1000 images to train (use ImageNet pretrained model ResNet-101) and test only in (720, 1280) scale.
Authors: Zijian Zhou, Zhizhong Su, Shuangping Huang
Description: This result is just use 1000 images to train (use ImageNet pretrained model ResNet-101) and test only in (720, 1280) scale.