YOLO / text detection - need help

[Anna Muller]

Hi,

Has anybody implemented text detection using YOLO or some other similar software? I wanted to tag screenshots of tiktok videos to extract text from certain areas (using VoTT), pass that into YOLOv3, and then alter the image to make the text more readable given the noisy background. 

However, I am stuck on the YOLOv3 part - does anyone have any experience with this/know of a simple distribution of YOLO to use? Or can give me any pointers in general?

I would greatly appreciate it

Thank you

[Zdenko Podobny]

Hello,

something like this I have on my "try todo list". 

As far as I understand you need to train YOLO for text detections:

https://github.com/Neerajj9/Text-Detection-using-Yolo-Algorithm-in-keras-tensorflow

https://towardsdatascience.com/object-detection-on-newspaper-images-using-yolov3-85acfa563080

https://medium.com/saarthi-ai/how-to-build-your-own-ocr-a5bb91b622ba

One of the most interesting bookmarks from this area is IMO https://roboflow.com - the offer to annotate images with their tool [1] and run training there + export model (even to RB PI[2]). They also offer  synthetic dataset The Open Poetry Vision  for OCR tasks. So maybe with it you speed up training for your project.

[1] https://roboflow.com/annotate

[2] https://blog.roboflow.com/how-to-deploy-a-yolov8-model-to-a-raspberry-pi/

[3] https://public.roboflow.com/object-detection/open-poetry-vision

BR,

Zdenko

ut 31. 1. 2023 o 16:30 Anna Muller <amul...@nd.edu> napísal(a):

--
You received this message because you are subscribed to the Google Groups "tesseract-ocr" group.
To unsubscribe from this group and stop receiving emails from it, send an email to tesseract-oc...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/tesseract-ocr/07858222-caec-4213-84d7-5a81eab9c4acn%40googlegroups.com.