def extract_text(img_path, model_path): ocr = pyclstm.ClstmOcr() ocr.load(model_path) imgFile = Image.open(img_path) text = ocr.recognize(imgFile) text.encode('utf-8') chars = ocr.recognize_chars(imgFile) prob = 1 index = 0 x = 0 #print text if (text.find(u':') != -1 and text.index(u':') < 5): index = text.index(u':') + 1 if (text.find(u' ') != -1 and (text.index(u' ') <= 3)): if (len(text) > text.index(u' ') + 1): index = text.index(u' ') + 1 for ind, j in enumerate(chars): #print j if ind >= index: prob *= j.confidence if j.char == u':': if j.x_position > x: x = j.x_position #return text[index:], prob, index return text, prob, index, x
def extract_text(img_path, model_path): ocr = pyclstm.ClstmOcr() ocr.load(model_path) imgFile = Image.open(img_path) text = ocr.recognize(imgFile) text.encode('utf-8') chars = ocr.recognize_chars(imgFile) prob = 1 index = 0 # print text if text.find(u':') != -1 and text.index(u':') < 3: index = text.index(u':') + 1 if text.find(u' ') != -1 and (text.index(u' ') <= 3): if (len(text) > text.index(u' ') + 1): index = text.index(u' ') + 1 for ind, j in enumerate(chars): if ind >= index: prob *= j.confidence return text, prob
def extract_text(img_path, model_path): ocr = pyclstm.ClstmOcr() ocr.load(model_path) imgFile = Image.open(img_path) text = ocr.recognize(imgFile) #print text text.encode('utf-8') chars = ocr.recognize_chars(imgFile) prob = 1 index = 0 #print text # if(text.find(u':') != -1 and text.index(u':') < 5): # index = text.index(u':')+1 # if(text.find(u' ') != -1 and (text.index(u' ') <= 3)): # if(len(text)>text.index(u' ')+1): # index = text.index(u' ')+1 for ind, j in enumerate(chars): #print j #print j if ind >= 0: prob *= j.confidence #print index return text[0:], prob, 0