Support for .heic file format

[datatalking]

What would be the process for extending support for pytesseract being able to work with .heic file types?

Essentially we are batch renaming the filetype to .jpg but they are not always supported.

I read through a bit of https://tesseract-ocr.github.io and did not see references to this file type for lossless files.

This is for a public works project using archived public documents going back over 200 years so having access to this filetype would be not only a public service but saving taxpayers money.

[bozhodimitrov]

Hi @datatalking,
I'm not sure if tesseract itself support this file format?
But it looks that there are lots of options suggested on StackOverflow: How to work with HEIC image file types in Python.
Very Interesting and exotic image format by the way - my assumption is that the pre-processing will be required anyways.

[datatalking]

Hi @int3l,
For argument's sake where might we patch/add the feature into pytesseract?

[bozhodimitrov]

One possibility is to change the prepare function. But it will require specific conversion for this format into one that is supported by Tesseract itself.

[datatalking]

@int3l What are the differences in the formats? As it stands now I have to convert thousands of them to jpg or other formats so I am curious what format Tesseract supports.

[bozhodimitrov]

Tesseract itself uses the Leptopnica image library -- here is unofficial list of supported formats: Leptopnica Image I/O formats