I'm experimenting with Apple's Vision framework to detect the location of characters (letters, numbers, etc). Why can't I get the VisionBasics demo project to detect text in images with less than 3 digits?
I've already tried binarizing the image by reducing saturation and increasing contrast. I even tried inverting the black and white portions, but it didn't improve the results. The 2-digit images are approximately 28x24 pixels.
Link to Xcode Project:
https://docs-assets.developer.apple.com/published/422afa2a8c/DetectingObjectsInStillImages.zip
Please also see the unanswered question posted at: Vision, VNDetectTextRectanglesRequest - can't recognize single number as region