Traditional Multilingual Text Recognition (MLTR) usually targets a fixed...
Text irregularities pose significant challenges to scene text recognizer...
Dominant scene text recognition models commonly contain two building blo...
The attention-based encoder-decoder framework is becoming popular in sce...
share