def __init__(self):
use_pdserving = False
use_zero_copy_run = False
rec_image_shape = "3, 32, 320"
rec_char_type = "ch"
rec_batch_num = 6
rec_algorithm = "CRNN"
max_text_length = 25
rec_char_dict_path = "./src/ai/ocr_paddle/ppocr/ppocr_keys_v1.txt"
use_space_char = True
if use_pdserving is False:
self.predictor, self.input_tensor, self.output_tensors =\
utility.create_predictor( mode="rec")
self.use_zero_copy_run = use_zero_copy_run
self.rec_image_shape = [int(v) for v in rec_image_shape.split(",")]
self.character_type = rec_char_type
self.rec_batch_num = rec_batch_num
self.rec_algorithm = rec_algorithm
self.text_len = max_text_length
char_ops_params = {
"character_type": rec_char_type,
"character_dict_path": rec_char_dict_path,
"use_space_char": use_space_char,
"max_text_length": max_text_length
}
if self.rec_algorithm in ["CRNN", "Rosetta", "STAR-Net"]:
char_ops_params['loss_type'] = 'ctc'
self.loss_type = 'ctc'
elif self.rec_algorithm == "RARE":
char_ops_params['loss_type'] = 'attention'
self.loss_type = 'attention'
elif self.rec_algorithm == "SRN":
char_ops_params['loss_type'] = 'srn'
self.loss_type = 'srn'
self.char_ops = CharacterOps(char_ops_params)
def __init__(self, args):
self.predictor, self.input_tensor, self.output_tensors =\
utility.create_predictor(args, mode="rec")
image_shape = [int(v) for v in args.rec_image_shape.split(",")]
self.rec_image_shape = image_shape
char_ops_params = {}
char_ops_params["character_type"] = args.rec_char_type
char_ops_params["character_dict_path"] = args.rec_char_dict_path
char_ops_params['loss_type'] = 'ctc'
self.char_ops = CharacterOps(char_ops_params)
def __init__(self):
max_side_len = config.det_max_side_len
self.det_algorithm = config.det_algorithm
preprocess_params = {'max_side_len': max_side_len}
postprocess_params = {}
self.preprocess_op = db_preprocess.DBProcessTest(preprocess_params)
postprocess_params["thresh"] = config.det_db_thresh
postprocess_params["box_thresh"] = config.det_db_box_thresh
postprocess_params["max_candidates"] = 1000
postprocess_params["unclip_ratio"] = config.det_db_unclip_ratio
self.postprocess_op = db_postprocess.DBPostProcess(postprocess_params)
self.predictor, self.input_tensor, self.output_tensors =\
utility.create_predictor(mode="detect")
def get_text_detector(
det_model_dir="./paddle_detector/inference/ch_ppocr_mobile_v2.0_det_infer",
cls_model_dir="./paddle_detector/inference/ch_ppocr_mobile_v2.0_cls_infer",
):
args = TextDetectorConfig()
args.det_model_dir = det_model_dir
args.cls_model_dir = cls_model_dir
args.use_angle_cls = True
args.use_gpu = False
text_detector = TextDetector(args)
# Patch detector to prevent memory leak
predictor, input_tensor, output_tensors = create_predictor(
args, "det", logger)
text_detector.predictor = predictor
text_detector.input_tensor = input_tensor
text_detector.output_tensors = output_tensors
return text_detector
def __init__(self):
self.predictor, self.input_tensor, self.output_tensors =\
utility.create_predictor(mode="rec")
image_shape = [int(v) for v in config.rec_image_shape.split(",")]
self.rec_image_shape = image_shape
self.character_type = config.rec_char_type
self.rec_batch_num = config.rec_batch_num
self.rec_algorithm = config.rec_algorithm
char_ops_params = {}
char_ops_params["character_type"] = config.rec_char_type
char_ops_params["character_dict_path"] = config.rec_char_dict_path
if self.rec_algorithm != "RARE":
char_ops_params['loss_type'] = 'ctc'
self.loss_type = 'ctc'
else:
char_ops_params['loss_type'] = 'attention'
self.loss_type = 'attention'
self.char_ops = character.CharacterOps(char_ops_params)
def __init__(self, args):
self.predictor, self.input_tensor, self.output_tensors =\
utility.create_predictor(args, mode="rec")
self.rec_image_shape = [
int(v) for v in args.rec_image_shape.split(",")
]
self.character_type = args.rec_char_type
self.rec_batch_num = args.rec_batch_num
self.rec_algorithm = args.rec_algorithm
char_ops_params = {
"character_type": args.rec_char_type,
"character_dict_path": args.rec_char_dict_path,
"use_space_char": args.use_space_char
}
if self.rec_algorithm != "RARE":
char_ops_params['loss_type'] = 'ctc'
self.loss_type = 'ctc'
else:
char_ops_params['loss_type'] = 'attention'
self.loss_type = 'attention'
self.char_ops = CharacterOps(char_ops_params)
def __init__(self, args):
max_side_len = args.det_max_side_len
self.det_algorithm = args.det_algorithm
preprocess_params = {'max_side_len': max_side_len}
postprocess_params = {}
if self.det_algorithm == "DB":
self.preprocess_op = DBProcessTest(preprocess_params)
postprocess_params["thresh"] = args.det_db_thresh
postprocess_params["box_thresh"] = args.det_db_box_thresh
postprocess_params["max_candidates"] = 1000
self.postprocess_op = DBPostProcess(postprocess_params)
elif self.det_algorithm == "EAST":
self.preprocess_op = EASTProcessTest(preprocess_params)
postprocess_params["score_thresh"] = args.det_east_score_thresh
postprocess_params["cover_thresh"] = args.det_east_cover_thresh
postprocess_params["nms_thresh"] = args.det_east_nms_thresh
self.postprocess_op = EASTPostPocess(postprocess_params)
else:
logger.info("unknown det_algorithm:{}".format(self.det_algorithm))
sys.exit(0)
self.predictor, self.input_tensor, self.output_tensors =\
utility.create_predictor(args, mode="det")