About evaluating the possibilities of extracting text from image diagrams with complex background
not conclusive, OCR not possible and SVg require custom cleaning details in gpt4/readme.md
- https://github.com/tesseract-ocr/tesseract
- https://tesseract-ocr.github.io/tessdoc/Command-Line-Usage.html#simplest-invocation-to-ocr-an-image
example command
tesseract test_slide.png test_slide --psm 11 -l eng pdf
help output see tessteract-help.md