强制tesseract-ocr识别单个字符？

Question

    api = tesseract.TessBaseAPI()
    api.SetOutputName("output")

    api.Init(".","eng",tesseract.OEM_DEFAULT)
    api.SetVariable("tessedit_char_whitelist", "0123456789")
    api.SetVariable("tessedit_pageseg_mode", "7")

    pixImage = tesseract.pixRead('img.jpg')

    api.SetImage(pixImage)

    outText = api.GetUTF8Text()
    answer = outText.replace("\n", "").replace(" ", "")

    print answer

上面的图像，即＆＃39; 1＆＃39;，被识别为＆＃39; 47＆＃39;通过tesseract已经设置为api.SetVariable("tessedit_pageseg_mode", "7")。如何强制它识别图像中的单个字符？

P.S。：python 2.7.3 + tesseract 3.0.2