1 paper
Minghui Liao, Jian Zhang, Zhaoyi Wan +5
Inspired by speech recognition, recent state-of-the-art algorithms mostly consider scene text recognition as a sequence prediction problem. Though achieving excellent performance,…