I know certain image processing effects can enhance OCR reading of text and I need results to be quite accurate for my project.

Should I add a process of image clarity through increased contrast and de-noise etc. before sending my image to the Google Vision API or does Google do that already? Concerned about us doubling up on something like contrast and deep-frying my image.

I can’t find any technical information from this online or from Google support contacts, but I wouldn’t be surprised if their vision OCR method does a process of image clarity for text readability.

I’ve only tried sending normal images and get decent results but they could be better – I haven’t added my own effects yet.