predict_det.py 11.8 KB
Newer Older
L
LDOUBLEV 已提交
1 2 3 4 5 6 7 8 9 10 11 12 13
# Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
L
LDOUBLEV 已提交
14 15
import os
import sys
W
WenmuZhou 已提交
16

17
__dir__ = os.path.dirname(os.path.abspath(__file__))
L
LDOUBLEV 已提交
18
sys.path.append(__dir__)
littletomatodonkey's avatar
littletomatodonkey 已提交
19
sys.path.insert(0, os.path.abspath(os.path.join(__dir__, '../..')))
L
LDOUBLEV 已提交
20

L
LDOUBLEV 已提交
21 22
os.environ["FLAGS_allocator_strategy"] = 'auto_growth'

23 24 25 26
import cv2
import numpy as np
import time
import sys
L
fix  
LDOUBLEV 已提交
27
from scipy.spatial import distance as dist
28

L
LDOUBLEV 已提交
29
import tools.infer.utility as utility
W
WenmuZhou 已提交
30
from ppocr.utils.logging import get_logger
L
LDOUBLEV 已提交
31
from ppocr.utils.utility import get_image_file_list, check_and_read_gif
W
WenmuZhou 已提交
32 33
from ppocr.data import create_operators, transform
from ppocr.postprocess import build_post_process
L
LDOUBLEV 已提交
34
import json
W
WenmuZhou 已提交
35 36
logger = get_logger()

L
LDOUBLEV 已提交
37 38 39

class TextDetector(object):
    def __init__(self, args):
L
LDOUBLEV 已提交
40
        self.args = args
L
LDOUBLEV 已提交
41
        self.det_algorithm = args.det_algorithm
T
tink2123 已提交
42
        self.use_onnx = args.use_onnx
M
MissPenguin 已提交
43
        pre_process_list = [{
44 45
            'DetResizeForTest': {
                'limit_side_len': args.det_limit_side_len,
W
WenmuZhou 已提交
46
                'limit_type': args.det_limit_type,
47
            }
M
MissPenguin 已提交
48 49 50 51 52 53 54 55 56 57 58 59 60 61
        }, {
            'NormalizeImage': {
                'std': [0.229, 0.224, 0.225],
                'mean': [0.485, 0.456, 0.406],
                'scale': '1./255.',
                'order': 'hwc'
            }
        }, {
            'ToCHWImage': None
        }, {
            'KeepKeys': {
                'keep_keys': ['image', 'shape']
            }
        }]
L
LDOUBLEV 已提交
62 63
        postprocess_params = {}
        if self.det_algorithm == "DB":
W
WenmuZhou 已提交
64
            postprocess_params['name'] = 'DBPostProcess'
L
LDOUBLEV 已提交
65 66 67
            postprocess_params["thresh"] = args.det_db_thresh
            postprocess_params["box_thresh"] = args.det_db_box_thresh
            postprocess_params["max_candidates"] = 1000
68
            postprocess_params["unclip_ratio"] = args.det_db_unclip_ratio
L
LDOUBLEV 已提交
69
            postprocess_params["use_dilation"] = args.use_dilation
littletomatodonkey's avatar
littletomatodonkey 已提交
70
            postprocess_params["score_mode"] = args.det_db_score_mode
M
MissPenguin 已提交
71
        elif self.det_algorithm == "EAST":
W
WenmuZhou 已提交
72
            postprocess_params['name'] = 'EASTPostProcess'
M
MissPenguin 已提交
73 74 75 76
            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
        elif self.det_algorithm == "SAST":
M
MissPenguin 已提交
77
            pre_process_list[0] = {
W
WenmuZhou 已提交
78 79 80
                'DetResizeForTest': {
                    'resize_long': args.det_limit_side_len
                }
M
MissPenguin 已提交
81
            }
W
WenmuZhou 已提交
82
            postprocess_params['name'] = 'SASTPostProcess'
M
MissPenguin 已提交
83 84 85 86 87 88 89 90 91 92 93
            postprocess_params["score_thresh"] = args.det_sast_score_thresh
            postprocess_params["nms_thresh"] = args.det_sast_nms_thresh
            self.det_sast_polygon = args.det_sast_polygon
            if self.det_sast_polygon:
                postprocess_params["sample_pts_num"] = 6
                postprocess_params["expand_scale"] = 1.2
                postprocess_params["shrink_ratio_of_width"] = 0.2
            else:
                postprocess_params["sample_pts_num"] = 2
                postprocess_params["expand_scale"] = 1.0
                postprocess_params["shrink_ratio_of_width"] = 0.3
W
WenmuZhou 已提交
94 95 96 97 98 99 100 101
        elif self.det_algorithm == "PSE":
            postprocess_params['name'] = 'PSEPostProcess'
            postprocess_params["thresh"] = args.det_pse_thresh
            postprocess_params["box_thresh"] = args.det_pse_box_thresh
            postprocess_params["min_area"] = args.det_pse_min_area
            postprocess_params["box_type"] = args.det_pse_box_type
            postprocess_params["scale"] = args.det_pse_scale
            self.det_pse_box_type = args.det_pse_box_type
文幕地方's avatar
文幕地方 已提交
102 103 104 105 106 107 108 109 110 111 112 113
        elif self.det_algorithm == "FCE":
            pre_process_list[0] = {
                'DetResizeForTest': {
                    'rescale_img': [1080, 736]
                }
            }
            postprocess_params['name'] = 'FCEPostProcess'
            postprocess_params["scales"] = args.scales
            postprocess_params["alpha"] = args.alpha
            postprocess_params["beta"] = args.beta
            postprocess_params["fourier_degree"] = args.fourier_degree
            postprocess_params["box_type"] = args.det_fce_box_type
L
LDOUBLEV 已提交
114 115 116
        else:
            logger.info("unknown det_algorithm:{}".format(self.det_algorithm))
            sys.exit(0)
117

W
WenmuZhou 已提交
118 119
        self.preprocess_op = create_operators(pre_process_list)
        self.postprocess_op = build_post_process(postprocess_params)
L
LDOUBLEV 已提交
120 121 122
        self.predictor, self.input_tensor, self.output_tensors, self.config = utility.create_predictor(
            args, 'det', logger)

123 124 125 126 127 128 129 130 131 132
        if self.use_onnx:
            img_h, img_w = self.input_tensor.shape[2:]
            if img_h is not None and img_w is not None and img_h > 0 and img_w > 0:
                pre_process_list[0] = {
                    'DetResizeForTest': {
                        'image_shape': [img_h, img_w]
                    }
                }
        self.preprocess_op = create_operators(pre_process_list)

D
Double_V 已提交
133
        if args.benchmark:
D
Double_V 已提交
134
            import auto_log
D
Double_V 已提交
135
            pid = os.getpid()
L
LDOUBLEV 已提交
136
            gpu_id = utility.get_infer_gpuid()
D
Double_V 已提交
137 138 139 140 141
            self.autolog = auto_log.AutoLogger(
                model_name="det",
                model_precision=args.precision,
                batch_size=1,
                data_shape="dynamic",
L
LDOUBLEV 已提交
142
                save_path=None,
D
Double_V 已提交
143 144 145
                inference_config=self.config,
                pids=pid,
                process_name=None,
146
                gpu_ids=gpu_id if args.use_gpu else None,
D
Double_V 已提交
147 148 149
                time_keys=[
                    'preprocess_time', 'inference_time', 'postprocess_time'
                ],
150
                warmup=2,
L
LDOUBLEV 已提交
151
                logger=logger)
L
LDOUBLEV 已提交
152

L
LDOUBLEV 已提交
153
    def order_points_clockwise(self, pts):
L
fix  
LDOUBLEV 已提交
154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169
        """
        refer to :https://github.com/PyImageSearch/imutils/blob/9f740a53bcc2ed7eba2558afed8b4c17fd8a1d4c/imutils/perspective.py#L9
        """
        # sort the points based on their x-coordinates
        xSorted = pts[np.argsort(pts[:, 0]), :]

        leftMost = xSorted[:2, :]
        rightMost = xSorted[2:, :]

        leftMost = leftMost[np.argsort(leftMost[:, 1]), :]
        (tl, bl) = leftMost

        D = dist.cdist(tl[np.newaxis], rightMost, "euclidean")[0]
        (br, tr) = rightMost[np.argsort(D)[::-1], :]

        return np.array([tl, tr, br, bl], dtype="float32")
文幕地方's avatar
文幕地方 已提交
170

D
dyning 已提交
171
    def clip_det_res(self, points, img_height, img_width):
172
        for pno in range(points.shape[0]):
D
dyning 已提交
173 174
            points[pno, 0] = int(min(max(points[pno, 0], 0), img_width - 1))
            points[pno, 1] = int(min(max(points[pno, 1], 0), img_height - 1))
L
LDOUBLEV 已提交
175 176 177 178 179 180 181
        return points

    def filter_tag_det_res(self, dt_boxes, image_shape):
        img_height, img_width = image_shape[0:2]
        dt_boxes_new = []
        for box in dt_boxes:
            box = self.order_points_clockwise(box)
D
dyning 已提交
182
            box = self.clip_det_res(box, img_height, img_width)
L
LDOUBLEV 已提交
183 184
            rect_width = int(np.linalg.norm(box[0] - box[1]))
            rect_height = int(np.linalg.norm(box[0] - box[3]))
M
MissPenguin 已提交
185
            if rect_width <= 3 or rect_height <= 3:
L
LDOUBLEV 已提交
186 187 188 189 190
                continue
            dt_boxes_new.append(box)
        dt_boxes = np.array(dt_boxes_new)
        return dt_boxes

191 192 193 194 195 196 197 198
    def filter_tag_det_res_only_clip(self, dt_boxes, image_shape):
        img_height, img_width = image_shape[0:2]
        dt_boxes_new = []
        for box in dt_boxes:
            box = self.clip_det_res(box, img_height, img_width)
            dt_boxes_new.append(box)
        dt_boxes = np.array(dt_boxes_new)
        return dt_boxes
199

L
LDOUBLEV 已提交
200 201
    def __call__(self, img):
        ori_im = img.copy()
W
WenmuZhou 已提交
202
        data = {'image': img}
L
LDOUBLEV 已提交
203 204

        st = time.time()
L
LDOUBLEV 已提交
205

littletomatodonkey's avatar
littletomatodonkey 已提交
206
        if self.args.benchmark:
D
Double_V 已提交
207
            self.autolog.times.start()
L
LDOUBLEV 已提交
208

W
WenmuZhou 已提交
209 210 211
        data = transform(data, self.preprocess_op)
        img, shape_list = data
        if img is None:
L
LDOUBLEV 已提交
212
            return None, 0
W
WenmuZhou 已提交
213 214
        img = np.expand_dims(img, axis=0)
        shape_list = np.expand_dims(shape_list, axis=0)
215
        img = img.copy()
L
LDOUBLEV 已提交
216

littletomatodonkey's avatar
littletomatodonkey 已提交
217
        if self.args.benchmark:
D
Double_V 已提交
218
            self.autolog.times.stamp()
T
tink2123 已提交
219 220 221 222 223 224 225 226 227 228 229 230 231
        if self.use_onnx:
            input_dict = {}
            input_dict[self.input_tensor.name] = img
            outputs = self.predictor.run(self.output_tensors, input_dict)
        else:
            self.input_tensor.copy_from_cpu(img)
            self.predictor.run()
            outputs = []
            for output_tensor in self.output_tensors:
                output = output_tensor.copy_to_cpu()
                outputs.append(output)
            if self.args.benchmark:
                self.autolog.times.stamp()
L
LDOUBLEV 已提交
232

M
MissPenguin 已提交
233 234 235 236 237 238 239 240 241
        preds = {}
        if self.det_algorithm == "EAST":
            preds['f_geo'] = outputs[0]
            preds['f_score'] = outputs[1]
        elif self.det_algorithm == 'SAST':
            preds['f_border'] = outputs[0]
            preds['f_score'] = outputs[1]
            preds['f_tco'] = outputs[2]
            preds['f_tvo'] = outputs[3]
W
WenmuZhou 已提交
242
        elif self.det_algorithm in ['DB', 'PSE']:
W
WenmuZhou 已提交
243
            preds['maps'] = outputs[0]
文幕地方's avatar
文幕地方 已提交
244 245 246
        elif self.det_algorithm == 'FCE':
            for i, output in enumerate(outputs):
                preds['level_{}'.format(i)] = output
W
WenmuZhou 已提交
247 248
        else:
            raise NotImplementedError
L
LDOUBLEV 已提交
249

L
LDOUBLEV 已提交
250
        #self.predictor.try_shrink_memory()
W
WenmuZhou 已提交
251 252
        post_result = self.postprocess_op(preds, shape_list)
        dt_boxes = post_result[0]['points']
文幕地方's avatar
文幕地方 已提交
253 254 255
        if (self.det_algorithm == "SAST" and self.det_sast_polygon) or (
                self.det_algorithm in ["PSE", "FCE"] and
                self.postprocess_op.box_type == 'poly'):
W
WenmuZhou 已提交
256
            dt_boxes = self.filter_tag_det_res_only_clip(dt_boxes, ori_im.shape)
M
MissPenguin 已提交
257 258
        else:
            dt_boxes = self.filter_tag_det_res(dt_boxes, ori_im.shape)
L
LDOUBLEV 已提交
259

littletomatodonkey's avatar
littletomatodonkey 已提交
260
        if self.args.benchmark:
D
Double_V 已提交
261
            self.autolog.times.end(stamp=True)
L
LDOUBLEV 已提交
262 263
        et = time.time()
        return dt_boxes, et - st
L
LDOUBLEV 已提交
264 265 266 267


if __name__ == "__main__":
    args = utility.parse_args()
L
LDOUBLEV 已提交
268
    image_file_list = get_image_file_list(args.image_dir)
L
LDOUBLEV 已提交
269 270 271
    text_detector = TextDetector(args)
    count = 0
    total_time = 0
littletomatodonkey's avatar
littletomatodonkey 已提交
272
    draw_img_save = "./inference_results"
L
LDOUBLEV 已提交
273

L
LDOUBLEV 已提交
274 275
    if args.warmup:
        img = np.random.uniform(0, 255, [640, 640, 3]).astype(np.uint8)
276
        for i in range(2):
L
LDOUBLEV 已提交
277 278
            res = text_detector(img)

littletomatodonkey's avatar
littletomatodonkey 已提交
279 280
    if not os.path.exists(draw_img_save):
        os.makedirs(draw_img_save)
L
LDOUBLEV 已提交
281
    save_results = []
L
LDOUBLEV 已提交
282
    for image_file in image_file_list:
L
LDOUBLEV 已提交
283 284 285
        img, flag = check_and_read_gif(image_file)
        if not flag:
            img = cv2.imread(image_file)
L
LDOUBLEV 已提交
286 287 288
        if img is None:
            logger.info("error in loading image:{}".format(image_file))
            continue
L
LDOUBLEV 已提交
289 290 291
        st = time.time()
        dt_boxes, _ = text_detector(img)
        elapse = time.time() - st
L
LDOUBLEV 已提交
292 293 294
        if count > 0:
            total_time += elapse
        count += 1
L
LDOUBLEV 已提交
295
        save_pred = os.path.basename(image_file) + "\t" + str(
文幕地方's avatar
文幕地方 已提交
296
            json.dumps([x.tolist() for x in dt_boxes])) + "\n"
L
LDOUBLEV 已提交
297 298
        save_results.append(save_pred)
        logger.info(save_pred)
L
fix log  
LDOUBLEV 已提交
299
        logger.info("The predict time of {}: {}".format(image_file, elapse))
D
dyning 已提交
300
        src_im = utility.draw_text_det_res(dt_boxes, image_file)
W
WenmuZhou 已提交
301
        img_name_pure = os.path.split(image_file)[-1]
W
WenmuZhou 已提交
302 303
        img_path = os.path.join(draw_img_save,
                                "det_res_{}".format(img_name_pure))
L
LDOUBLEV 已提交
304
        cv2.imwrite(img_path, src_im)
W
WenmuZhou 已提交
305
        logger.info("The visualized image saved in {}".format(img_path))
L
LDOUBLEV 已提交
306

L
LDOUBLEV 已提交
307 308 309
    with open(os.path.join(draw_img_save, "det_results.txt"), 'w') as f:
        f.writelines(save_results)
        f.close()
D
Double_V 已提交
310 311
    if args.benchmark:
        text_detector.autolog.report()