Open ishandutta2007 opened 3 months ago
The text detector model is detecting that logo as text. You need to find tune the text detector model (CRAFT) so it doesn't correctly detects the text regions.
What reader.readtext()
does is that it first inference using the text detections model to extract the text regions, then the recognizer model to recognize the characters/texts
Train your own CRAFT model - https://github.com/JaidedAI/EasyOCR/tree/master/trainer/craft
@darrenchang I am looking for an easier work around to reject it based on say colour or size .
I have an image like this it is always reading like this @61.23
I just wanted to read like 61.23