ocrModel = ocrModelTorchDense
else:
ocrModel = ocrModelTorchEng
alphabet = alphabetEnglish
LSTMFLAG = True
elif ocrFlag == 'opencv':
from crnn.network_dnn import CRNN
ocrModel = ocrModelOpencv
alphabet = alphabetChinese
else:
print("err,ocr engine in keras\opencv\darknet")
nclass = len(alphabet) + 1
if ocrFlag == 'opencv':
crnn = CRNN(alphabet=alphabet)
else:
crnn = CRNN(32,
1,
nclass,
256,
leakyRelu=False,
lstmFlag=LSTMFLAG,
GPU=GPU,
alphabet=alphabet)
if os.path.exists(ocrModel):
crnn.load_weights(ocrModel)
else:
print("download model or tranform model with tools!")
ocr = crnn.predict_job
def crnnSource():
alphabet = keys.alphabetChinese##中英文模型
converter = strLabelConverter(alphabet)
model = CRNN(32, 1, len(alphabet)+1, 256, 1,lstmFlag=LSTMFLAG)
model.load_weights(ocrModelKeras)
return model,converter
# coding:utf-8
from crnn.utils import strLabelConverter, resizeNormalize
from crnn.network_keras import keras_crnn as CRNN
import tensorflow as tf
from crnn import keys
from config import ocrModelKeras
import numpy as np
graph = tf.get_default_graph()
alphabet = keys.alphabetChinese
model = CRNN(32, 1, len(alphabet) + 1, 256, 1, lstmFlag=False)
model.load_weights(ocrModelKeras)
export_path = "models/crnn/1"
with tf.keras.backend.get_session() as sess:
out = model.output
keys = [0]
values = [" "]
for idx, char in enumerate(alphabet):
keys.append(idx + 1)
values.append(char)
table = tf.contrib.lookup.HashTable(
tf.contrib.lookup.KeyValueTensorInitializer(keys, values), "ç")