search

sirfz / tesserocr

A Python wrapper for the tesseract-ocr API
MIT License
2.02k stars 254 forks source link

word by word result different to entire document #180

Closed ckalas closed 5 years ago

ckalas commented 5 years ago 
from PIL import Image
from tesserocr import PyTessBaseAPI, RIL

image = Image.open('/input_tiffs/one.tif')

with PyTessBaseAPI() as api:

    api.SetImage(image)
    ALL = api.GetUTF8Text()
    boxes = api.GetComponentImages(RIL.WORD, True)
    print(f'Found {len(boxes)} textline image components.')
    words = []
    for i, (im, box, _, _) in enumerate(boxes):
        # im is a PIL image object
        # box is a dict with x, y, w and h keys
        api.SetRectangle(box['x'], box['y'], box['w'], box['h'])
        ocrResult = api.GetUTF8Text()
        conf = api.MeanTextConf()
        words.append(ocrResult)

    # ALL is different to words, with missing and incorrect words

Why do I get different (incorrect) body of text by combining the words at a word level than when calling GetUTF8Text() on the entire image?

sirfz commented 5 years ago

This is a tesseract-specific question rather than tessererocr which is just a wrapper around the tesseract API. You're better off asking on StackOverflow imho.