add text normalize

kotaro-kinoshita · Nov 28, 2024 · 3747cea · 3747cea
1 parent d71ebb9
commit 3747cea
Show file tree

Hide file tree

Showing 2 changed files with 4 additions and 1 deletion.
diff --git a/src/yomitoku/text_recognizer.py b/src/yomitoku/text_recognizer.py
@@ -2,6 +2,7 @@
 
 import numpy as np
 import torch
+import unicodedata
 from pydantic import conlist
 
 from .base import BaseModelCatalog, BaseModule, BaseSchema
@@ -72,6 +73,8 @@ def preprocess(self, img, polygons):
 
     def postprocess(self, p, points):
         pred, score = self.tokenizer.decode(p)
+        pred = [unicodedata.normalize("NFKC", x) for x in pred]
+
         directions = []
         for point in points:
             point = np.array(point)

diff --git a/uv.lock b/uv.lock