What preprocessing operations are performed by Tesseract OCR?

[Sirius Lee]

I couldn't find a detailed documentation and I don't feel browsing the source code. I want not to redo canny edge detection for example if it is already done by Tesseract engine. I need to know its steps before doing mine