add srn for dygraph

ppocr/postprocess/__init__.py

@@ -26,11 +26,12 @@ def build_post_process(config, global_config=None):
     from .db_postprocess import DBPostProcess
     from .east_postprocess import EASTPostProcess
     from .sast_postprocess import SASTPostProcess
-    from .rec_postprocess import CTCLabelDecode, AttnLabelDecode
+    from .rec_postprocess import CTCLabelDecode, AttnLabelDecode, SRNLabelDecode
     from .cls_postprocess import ClsPostProcess
 
     support_dict = [
-        'DBPostProcess', 'EASTPostProcess', 'SASTPostProcess', 'CTCLabelDecode', 'AttnLabelDecode', 'ClsPostProcess'
+        'DBPostProcess', 'EASTPostProcess', 'SASTPostProcess', 'CTCLabelDecode',
+        'AttnLabelDecode', 'ClsPostProcess', 'SRNLabelDecode'
     ]
 
     config = copy.deepcopy(config)

ppocr/postprocess/rec_postprocess.py

@@ -29,6 +29,9 @@ class BaseRecLabelDecode(object):
         assert character_type in support_character_type, "Only {} are supported now but get {}".format(
             support_character_type, character_type)
 
+        self.beg_str = "sos"
+        self.end_str = "eos"
+
         if character_type == "en":
             self.character_str = "0123456789abcdefghijklmnopqrstuvwxyz"
             dict_character = list(self.character_str)
@@ -104,7 +107,6 @@ class CTCLabelDecode(BaseRecLabelDecode):
     def __call__(self, preds, label=None, *args, **kwargs):
         if isinstance(preds, paddle.Tensor):
             preds = preds.numpy()
-
         preds_idx = preds.argmax(axis=2)
         preds_prob = preds.max(axis=2)
         text = self.decode(preds_idx, preds_prob)
@@ -153,3 +155,83 @@ class AttnLabelDecode(BaseRecLabelDecode):
             assert False, "unsupport type %s in get_beg_end_flag_idx" \
                           % beg_or_end
         return idx
+
+
+class SRNLabelDecode(BaseRecLabelDecode):
+    """ Convert between text-label and text-index """
+
+    def __init__(self,
+                 character_dict_path=None,
+                 character_type='en',
+                 use_space_char=False,
+                 **kwargs):
+        super(SRNLabelDecode, self).__init__(character_dict_path,
+                                             character_type, use_space_char)
+
+    def __call__(self, preds, label=None, *args, **kwargs):
+        pred = preds['predict']
+        char_num = len(self.character_str) + 2
+        if isinstance(pred, paddle.Tensor):
+            pred = pred.numpy()
+        pred = np.reshape(pred, [-1, char_num])
+
+        preds_idx = np.argmax(pred, axis=1)
+        preds_prob = np.max(pred, axis=1)
+
+        preds_idx = np.reshape(preds_idx, [-1, 25])
+
+        preds_prob = np.reshape(preds_prob, [-1, 25])
+
+        text = self.decode(preds_idx, preds_prob)
+
+        if label is None:
+            return text
+        label = self.decode(label, is_remove_duplicate=False)
+        return text, label
+
+    def decode(self, text_index, text_prob=None, is_remove_duplicate=True):
+        """ convert text-index into text-label. """
+        result_list = []
+        ignored_tokens = self.get_ignored_tokens()
+        batch_size = len(text_index)
+
+        for batch_idx in range(batch_size):
+            char_list = []
+            conf_list = []
+            for idx in range(len(text_index[batch_idx])):
+                if text_index[batch_idx][idx] in ignored_tokens:
+                    continue
+                if is_remove_duplicate:
+                    # only for predict
+                    if idx > 0 and text_index[batch_idx][idx - 1] == text_index[
+                            batch_idx][idx]:
+                        continue
+                char_list.append(self.character[int(text_index[batch_idx][
+                    idx])])
+                if text_prob is not None:
+                    conf_list.append(text_prob[batch_idx][idx])
+                else:
+                    conf_list.append(1)
+
+            text = ''.join(char_list)
+            result_list.append((text, np.mean(conf_list)))
+        return result_list
+
+    def add_special_char(self, dict_character):
+        dict_character = dict_character + [self.beg_str, self.end_str]
+        return dict_character
+
+    def get_ignored_tokens(self):
+        beg_idx = self.get_beg_end_flag_idx("beg")
+        end_idx = self.get_beg_end_flag_idx("end")
+        return [beg_idx, end_idx]
+
+    def get_beg_end_flag_idx(self, beg_or_end):
+        if beg_or_end == "beg":
+            idx = np.array(self.dict[self.beg_str])
+        elif beg_or_end == "end":
+            idx = np.array(self.dict[self.end_str])
+        else:
+            assert False, "unsupport type %s in get_beg_end_flag_idx" \
+                          % beg_or_end
+        return idx