yu20103983 / FOTS

FOTS tensorflow implementation, Include train and test, EAST+Rotate+CRNN. FOTS: Fast Oriented Text Spotting with a Unified Network
89 stars 35 forks source link

In the training of recognize, the text label is not mapping to the crop image? #20

Closed zh1049 closed 5 years ago

zh1049 commented 5 years ago

In the training of recognize, the text label is not mapping to the crop image? thanks