1 paper
Jiajun Wei, Hongjian Zhan, Xiao Tu +2
Employing a dictionary can efficiently rectify the deviation between the visual prediction and the ground truth in scene text recognition methods. However, the independence of the…