absence of ground truth for the generated dataset

Belval / TextRecognitionDataGenerator

A synthetic data generator for text recognition

MIT License

3.15k stars 943 forks source link

Open fardinkhanz opened 5 months ago

fardinkhanz commented 5 months ago

I am comparing few OCR models with the dataset but I need to generate ground truth for the data as well to use metrics like CER and WER.