numbers not properly recognised

[Swiftology]

Tesseract works fine (especially for word recognition) till I try to test it on a string of 14 or more digits. The result is inconsistent, getting only a few numbers right.
I used the code below:

                var tesseract:G8Tesseract = G8Tesseract(language:"eng")
                tesseract.engineMode = .TesseractCubeCombined
                tesseract.pageSegmentationMode = .Auto
                tesseract.maximumRecognitionTime = 60.0
                tesseract.image = newImage.g8_blackAndWhite()
                tesseract.recognize()
                let scannedText = tesseract.recognizedText

I am mainly using Calibri font. Could this be the reason?

[kevincon]

Unfortunately Tesseract alone is often not sufficient to consistently achieve high recognition accuracy. You can try some of the tips in our wiki: https://github.com/gali8/Tesseract-OCR-iOS/wiki/Tips-for-Improving-OCR-Results. In particular, try using GPUImage to preprocess the image before you give it to Tesseract.

[ws233]

@Swiftology, Tesseract doesn't support cube mode anymore.
pls, refer to #140
and tesseract-ocr/tesseract#40 for details.
Try default mode instead.

[Swiftology]

@kevincon I tried using GPUImage preprocessing but the results seem to be no better.

[Swiftology]

@ws233, I tried the default mode but the results are still the same, no improvements.

[ws233]

Your next step is training tesseract specifically to your font/fonts. Pls, refer to the documentation for more details.

[AndreaIOSDeveloper]

Is it possible to recognize only the numbers?