1 paper
Chenmin Yu, Liu Yu, Daiqing Wu +3
Modern visual object trackers show impressive results on general targets, yet their performance drops substantially when dealing with scene text. Although currently underexplored,…