Tesseract.js wraps an Emscripten port of the Tesseract OCR Engine
1–10 of 62 posts
Re: Tesseract.js wraps an Emscripten port of the Tesseract OCR Engine
#2This is a wrapper around a c++ codebase compiled with emscripten, so "pure" in the title doesn't really make sense.
Re: Tesseract.js wraps an Emscripten port of the Tesseract OCR Engine
#3Can it be considered "pure" when the project uses WASM ?
Re: Tesseract.js wraps an Emscripten port of the Tesseract OCR Engine
#4I imagine this is celebrating the V2 version release with WebAssembly? I believe it _used_ to be pure JS?
Re: Tesseract.js wraps an Emscripten port of the Tesseract OCR Engine
#5Can I do OCR all-in-browser with this, without involving any backend? Not much familiar with OCR accuracy metrics, how much accurate is Tesseract?
Re: Tesseract.js wraps an Emscripten port of the Tesseract OCR Engine
#6Can it be considered "pure" when the project uses WASM ?
I think people just mean 'no server side ML' when they say pure JS in this context.
Re: Tesseract.js wraps an Emscripten port of the Tesseract OCR Engine
#7Can I do OCR all-in-browser with this, without involving any backend? Not much familiar with OCR accuracy metrics, how much accurate is Tesseract?
Yes, this runs in your browser.
Re: Tesseract.js wraps an Emscripten port of the Tesseract OCR Engine
#8Can I do OCR all-in-browser with this, without involving any backend? Not much familiar with OCR accuracy metrics, how much accurate is Tesseract?
> Can I do OCR all-in-browser with this
Literally third sentence of the description:
> It works in the browser using webpack or plain script tags with a CDN and on the server with Node.js.
Re: Tesseract.js wraps an Emscripten port of the Tesseract OCR Engine
#9Can I do OCR all-in-browser with this, without involving any backend? Not much familiar with OCR accuracy metrics, how much accurate is Tesseract?
Tesseract works well with clean input, though in my experience it suffers greatly as soon as anything gets noisy.
Re: Tesseract.js wraps an Emscripten port of the Tesseract OCR Engine
#10No post body was provided.