mexican-government-report VS EasyOCR

Mexican Government Report Text Analysis using spaCy, NumPy, pandas, Matplotlib, Seaborn and geopandas.

This software was made possible due to EasyOCR (free) for Python. This OCR was incredibly easy to use, and extremely accurate for what I needed. A big thank you to the creators.

I've had good results with EasyOCR, much better than Tesseract. I agree with you, Tesseract has performed very poorly in my experience.

https://github.com/JaidedAI/EasyOCR

Currently the tool supports 2 different OCRs. PaddleOCR and EasyOCR. In both cases, the OCR has a specific model for Japanese characters. They work quite well, as long as the characters have clear contrast. The problem is that these OCR implement Torch, which makes the program very heavy. I am looking for a way to optimize this.

Try easyocr or Tesseract. Both are pretty easy to use and don't need much background in OpenCV.

EasyOCR: - Deployment: Easy as hell but difficult to train with personnalize data.

Process: I take an image. I take a screenshot of the first word I see. I register this screenshot. I give him the name 1.jpg,2.jpg,.... in a labels.csv, I write the name of the jpg and the corresponding text. I repeat... And I can use it to train https://github.com/JaidedAI/EasyOCR !

the repo with Python package https://github.com/JaidedAI/EasyOCR

website with online demo https://www.jaided.ai/easyocr/ (you can upload your images there)

Try running https://github.com/JaidedAI/EasyOCR, which reads all the text from a given image. Then loop over the words read and see if it reads your date properly or if it splits it into multiple words. Even if it does this, you can easily find the year (you might need to do some postprocessing on what the ocr reads, i.e transform l to 1). The nice thing is that you get the bounding boxes of where the tect was read so you can use some sort of postprocessing on bbox locations and words read to find the month amd day (i.e. I read 25 at the same y index as the yesr but the x index is to the left, so its probably the day). Good luck :)