Merge pull request #435 from LDOUBLEV/fixocr

read gif read func

ppocr/data/det/db_process.py

@@ -17,7 +17,7 @@ import cv2
 import numpy as np
 import json
 import sys
-from ppocr.utils.utility import initial_logger
+from ppocr.utils.utility import initial_logger, check_and_read_gif
 logger = initial_logger()
 
 from .data_augment import AugmentData
@@ -100,7 +100,9 @@ class DBProcessTrain(object):
 
     def __call__(self, label_infor):
         img_path, gt_label = self.convert_label_infor(label_infor)
-        imgvalue = cv2.imread(img_path)
+        imgvalue, flag = check_and_read_gif(img_path)
+        if not flag:
+            imgvalue = cv2.imread(img_path)
         if imgvalue is None:
             logger.info("{} does not exist!".format(img_path))
             return None

ppocr/utils/utility.py

@@ -15,6 +15,8 @@
 import logging
 import os
 import imghdr
+import cv2
+from paddle import fluid
 
 
 def initial_logger():
@@ -62,7 +64,7 @@ def get_image_file_list(img_file):
     if img_file is None or not os.path.exists(img_file):
         raise Exception("not found any img file in {}".format(img_file))
 
-    img_end = {'jpg', 'bmp', 'png', 'jpeg', 'rgb', 'tif', 'tiff'}
+    img_end = {'jpg', 'bmp', 'png', 'jpeg', 'rgb', 'tif', 'tiff', 'gif', 'GIF'}
     if os.path.isfile(img_file) and imghdr.what(img_file) in img_end:
         imgs_lists.append(img_file)
     elif os.path.isdir(img_file):
@@ -75,7 +77,18 @@ def get_image_file_list(img_file):
     return imgs_lists
 
 
-from paddle import fluid
+def check_and_read_gif(img_path):
+    if os.path.basename(img_path)[-3:] in ['gif', 'GIF']:
+        gif = cv2.VideoCapture(img_path)
+        ret, frame = gif.read()
+        if not ret:
+            logging.info("Cannot read {}. This gif image maybe corrupted.")
+            return None, False
+        if len(frame.shape) == 2 or frame.shape[-1] == 1:
+            frame = cv2.cvtColor(frame, cv2.COLOR_GRAY2RGB)
+        imgvalue = frame[:, :, ::-1]
+        return imgvalue, True
+    return None, False
 
 
 def create_multi_devices_program(program, loss_var_name):

tools/infer/predict_det.py

@@ -20,7 +20,7 @@ sys.path.append(os.path.join(__dir__, '../..'))
 import tools.infer.utility as utility
 from ppocr.utils.utility import initial_logger
 logger = initial_logger()
-from ppocr.utils.utility import get_image_file_list
+from ppocr.utils.utility import get_image_file_list, check_and_read_gif
 import cv2
 from ppocr.data.det.east_process import EASTProcessTest
 from ppocr.data.det.db_process import DBProcessTest
@@ -139,7 +139,9 @@ if __name__ == "__main__":
     if not os.path.exists(draw_img_save):
         os.makedirs(draw_img_save)
     for image_file in image_file_list:
-        img = cv2.imread(image_file)
+        img, flag = check_and_read_gif(image_file)
+        if not flag:
+            img = cv2.imread(image_file)
         if img is None:
             logger.info("error in loading image:{}".format(image_file))
             continue

tools/infer/predict_rec.py

@@ -20,7 +20,7 @@ sys.path.append(os.path.abspath(os.path.join(__dir__, '../..')))
 import tools.infer.utility as utility
 from ppocr.utils.utility import initial_logger
 logger = initial_logger()
-from ppocr.utils.utility import get_image_file_list
+from ppocr.utils.utility import get_image_file_list, check_and_read_gif
 import cv2
 import copy
 import numpy as np
@@ -153,7 +153,9 @@ def main(args):
     valid_image_file_list = []
     img_list = []
     for image_file in image_file_list:
-        img = cv2.imread(image_file, cv2.IMREAD_COLOR)
+        img, flag = check_and_read_gif(image_file)
+        if not flag:
+            img = cv2.imread(image_file)
         if img is None:
             logger.info("error in loading image:{}".format(image_file))
             continue

tools/infer/predict_system.py

@@ -27,7 +27,7 @@ import copy
 import numpy as np
 import math
 import time
-from ppocr.utils.utility import get_image_file_list
+from ppocr.utils.utility import get_image_file_list, check_and_read_gif
 from PIL import Image
 from tools.infer.utility import draw_ocr
 from tools.infer.utility import draw_ocr_box_txt
@@ -49,18 +49,23 @@ class TextSystem(object):
         points[:, 0] = points[:, 0] - left
         points[:, 1] = points[:, 1] - top
         '''
-        img_crop_width = int(max(np.linalg.norm(points[0] - points[1]),
-                                 np.linalg.norm(points[2] - points[3])))
-        img_crop_height = int(max(np.linalg.norm(points[0] - points[3]),
-                                  np.linalg.norm(points[1] - points[2])))
-        pts_std = np.float32([[0, 0],
-                              [img_crop_width, 0],
+        img_crop_width = int(
+            max(
+                np.linalg.norm(points[0] - points[1]),
+                np.linalg.norm(points[2] - points[3])))
+        img_crop_height = int(
+            max(
+                np.linalg.norm(points[0] - points[3]),
+                np.linalg.norm(points[1] - points[2])))
+        pts_std = np.float32([[0, 0], [img_crop_width, 0],
                               [img_crop_width, img_crop_height],
                               [0, img_crop_height]])
         M = cv2.getPerspectiveTransform(points, pts_std)
-        dst_img = cv2.warpPerspective(img, M, (img_crop_width, img_crop_height),
-                                      borderMode=cv2.BORDER_REPLICATE,
-                                      flags=cv2.INTER_CUBIC)
+        dst_img = cv2.warpPerspective(
+            img,
+            M, (img_crop_width, img_crop_height),
+            borderMode=cv2.BORDER_REPLICATE,
+            flags=cv2.INTER_CUBIC)
         dst_img_height, dst_img_width = dst_img.shape[0:2]
         if dst_img_height * 1.0 / dst_img_width >= 1.5:
             dst_img = np.rot90(dst_img)
@@ -119,25 +124,27 @@ def main(args):
     is_visualize = True
     tackle_img_num = 0
     for image_file in image_file_list:
-        img = cv2.imread(image_file)
+        img, flag = check_and_read_gif(image_file)
+        if not flag:
+            img = cv2.imread(image_file)
         if img is None:
             logger.info("error in loading image:{}".format(image_file))
             continue
         starttime = time.time()
-        tackle_img_num += 1	
-        if not args.use_gpu and args.enable_mkldnn and tackle_img_num % 30 == 0:	
+        tackle_img_num += 1
+        if not args.use_gpu and args.enable_mkldnn and tackle_img_num % 30 == 0:
             text_sys = TextSystem(args)
         dt_boxes, rec_res = text_sys(img)
         elapse = time.time() - starttime
         print("Predict time of %s: %.3fs" % (image_file, elapse))
+
+        drop_score = 0.5
         dt_num = len(dt_boxes)
-        dt_boxes_final = []
         for dno in range(dt_num):
             text, score = rec_res[dno]
-            if score >= 0.5:
+            if score >= drop_score:
                 text_str = "%s, %.3f" % (text, score)
                 print(text_str)
-                dt_boxes_final.append(dt_boxes[dno])
 
         if is_visualize:
             image = Image.fromarray(cv2.cvtColor(img, cv2.COLOR_BGR2RGB))
@@ -146,7 +153,12 @@ def main(args):
             scores = [rec_res[i][1] for i in range(len(rec_res))]
 
             draw_img = draw_ocr(
-                image, boxes, txts, scores, draw_txt=True, drop_score=0.5)
+                image,
+                boxes,
+                txts,
+                scores,
+                draw_txt=True,
+                drop_score=drop_score)
             draw_img_save = "./inference_results/"
             if not os.path.exists(draw_img_save):
                 os.makedirs(draw_img_save)