Make more improvements

paddleocr.py

@@ -637,7 +637,7 @@ def __init__(self, **kwargs):
     def ocr(self, img, det=True, rec=True, cls=True, bin=False, inv=False, alpha_color=(255, 255, 255)):
         """
         OCR with PaddleOCR
-
+        
         args:
             img: img for OCR, support ndarray, img_path and list or ndarray
             det: use text detection or not. If False, only rec will be exec. Default is True
@@ -832,7 +832,7 @@ def main():
                     outfile = args.output + '/' + img_name + '.txt'
                     with open(outfile,'w',encoding='utf-8') as f:
                         f.writelines(lines)
-
+                     
         elif args.type == 'structure':
             img, flag_gif, flag_pdf = check_and_read(img_path)
             if not flag_gif and not flag_pdf:

ppstructure/predict_system.py

@@ -123,13 +123,13 @@ def __call__(self, img, return_ocr_result_in_table=False, img_idx=0):
             # has problems with OCR recognition accuracy.
             #
             # To enhance the OCR recognition accuracy, we implement a patch fix
-            # that first detect all text regions by using the text_detector
-            # and then recognize the texts from the text regions (intersecting
-            # with the layout regions) by using the text_recognizer.
-            dt_boxes = []
+            # that first use text_system to detect and recognize all text information
+            # and then filter out relevant texts according to the layout regions.
+            text_res = None
             if self.text_system is not None:
-                dt_boxes, elapse = self.text_system.text_detector(img)
-                time_dict['det'] = elapse
+                text_res, ocr_time_dict = self._predict_text(img)
+                time_dict['det'] += ocr_time_dict['det']
+                time_dict['rec'] += ocr_time_dict['rec']
 
             res_list = []
             for region in layout_res:
@@ -152,10 +152,9 @@ def __call__(self, img, return_ocr_result_in_table=False, img_idx=0):
                         time_dict['det'] += table_time_dict['det']
                         time_dict['rec'] += table_time_dict['rec']
                 else:
-                    if self.text_system is not None:
-                        res, ocr_time_dict = self._predict_text(ori_im, roi_img, bbox, dt_boxes)
-                        time_dict['det'] += ocr_time_dict['det']
-                        time_dict['rec'] += ocr_time_dict['rec']
+                    if text_res is not None:
+                        # Filter the text results whose regions intersect with the current layout bbox.
+                        res = self._filter_text_res(text_res, bbox)
 
                 res_list.append({
                     'type': region['label'].lower(),
@@ -177,20 +176,8 @@ def __call__(self, img, return_ocr_result_in_table=False, img_idx=0):
 
         return None, None
 
-    def _predict_text(self, ori_im, roi_img, bbox, dt_boxes):
-        x1, y1, x2, y2 = bbox
-
-        if self.recovery:
-            wht_im = np.ones(ori_im.shape, dtype=ori_im.dtype)
-            wht_im[y1:y2, x1:x2, :] = roi_img
-            filter_boxes, filter_rec_res, ocr_time_dict = self.text_system(
-                wht_im)
-        else:
-            # Filter the text regions that intersect with the current bbox.
-            intersecting_dt_boxes = self._filter_boxes(dt_boxes, bbox)
-            # Recognize texts from these intersecting text regions.
-            filter_boxes, filter_rec_res, ocr_time_dict = self.text_system(
-                ori_im, dt_boxes=intersecting_dt_boxes)
+    def _predict_text(self, img):
+        filter_boxes, filter_rec_res, ocr_time_dict = self.text_system(img)
 
         # remove style char,
         # when using the recognition model trained on the PubtabNet dataset,
@@ -226,16 +213,14 @@ def _predict_text(self, ori_im, roi_img, bbox, dt_boxes):
                 })
         return res, ocr_time_dict
 
-    def _filter_boxes(self, dt_boxes, bbox):
-        boxes = []
-
-        for idx in range(len(dt_boxes)):
-            box = dt_boxes[idx]
+    def _filter_text_res(self, text_res, bbox):
+        res = []
+        for r in text_res:
+            box = r['text_region']
             rect = box[0][0], box[0][1], box[2][0], box[2][1]
             if self._has_intersection(bbox, rect):
-                boxes.append(box.tolist())
-
-        return np.array(boxes, np.float32).reshape((len(boxes), 4, 2))
+                res.append(r)
+        return res
 
     def _has_intersection(self, rect1, rect2):
         x_min1, y_min1, x_max1, y_max1 = rect1

tools/infer/predict_system.py

@@ -64,7 +64,7 @@ def draw_crop_rec_res(self, output_dir, img_crop_list, rec_res):
             logger.debug(f"{bno}, {rec_res[bno]}")
         self.crop_image_res_index += bbox_num
 
-    def __call__(self, img, cls=True, dt_boxes=None):
+    def __call__(self, img, cls=True):
         time_dict = {'det': 0, 'rec': 0, 'cls': 0, 'all': 0}
 
         if img is None:
@@ -73,11 +73,8 @@ def __call__(self, img, cls=True, dt_boxes=None):
 
         start = time.time()
         ori_im = img.copy()
-
-        elapse = 0
-        if dt_boxes is None:
-            dt_boxes, elapse = self.text_detector(img)
-            time_dict['det'] = elapse
+        dt_boxes, elapse = self.text_detector(img)
+        time_dict['det'] = elapse
 
         if dt_boxes is None:
             logger.debug("no dt_boxes found, elapsed : {}".format(elapse))