1 paper · 1 filter
Chuang Yang, Xu Han, Tao Han +3
Pursuing efficient text shape representations helps scene text detection models focus on compact foreground regions and optimize the contour reconstruction steps to simplify the wh…