From ab0b4e74744eac3593d8754688c1c144fcadca57 Mon Sep 17 00:00:00 2001
From: andyjpaddle <jiangkaitao@baidu.com>
Date: Tue, 2 Aug 2022 11:39:51 +0000
Subject: [PATCH] fix vl infer

---
 ppocr/data/imaug/rec_img_aug.py | 25 +++++++++++--------------
 tools/infer/predict_rec.py      |  1 +
 tools/infer_rec.py              |  1 -
 3 files changed, 12 insertions(+), 15 deletions(-)

diff --git a/ppocr/data/imaug/rec_img_aug.py b/ppocr/data/imaug/rec_img_aug.py
index ed3026a5..725b4b06 100644
--- a/ppocr/data/imaug/rec_img_aug.py
+++ b/ppocr/data/imaug/rec_img_aug.py
@@ -219,21 +219,18 @@ class VLRecResizeImg(object):
 
     def __call__(self, data):
         img = data['image']
-        if self.infer_mode and self.character_dict_path is not None:
-            norm_img, valid_ratio = resize_norm_img_chinese(img,
-                                                            self.image_shape)
+
+        imgC, imgH, imgW = self.image_shape
+        resized_image = cv2.resize(
+            img, (imgW, imgH), interpolation=cv2.INTER_LINEAR)
+        resized_w = imgW
+        resized_image = resized_image.astype('float32')
+        if self.image_shape[0] == 1:
+            resized_image = resized_image / 255
+            norm_img = resized_image[np.newaxis, :]
         else:
-            imgC, imgH, imgW = self.image_shape
-            resized_image = cv2.resize(
-                img, (imgW, imgH), interpolation=cv2.INTER_LINEAR)
-            resized_w = imgW
-            resized_image = resized_image.astype('float32')
-            if self.image_shape[0] == 1:
-                resized_image = resized_image / 255
-                norm_img = resized_image[np.newaxis, :]
-            else:
-                norm_img = resized_image.transpose((2, 0, 1)) / 255
-            valid_ratio = min(1.0, float(resized_w / imgW))
+            norm_img = resized_image.transpose((2, 0, 1)) / 255
+        valid_ratio = min(1.0, float(resized_w / imgW))
 
         data['image'] = norm_img
         data['valid_ratio'] = valid_ratio
diff --git a/tools/infer/predict_rec.py b/tools/infer/predict_rec.py
index 911f7511..e1d0d596 100755
--- a/tools/infer/predict_rec.py
+++ b/tools/infer/predict_rec.py
@@ -166,6 +166,7 @@ class TextRecognizer(object):
     def resize_norm_img_vl(self, img, image_shape):
 
         imgC, imgH, imgW = image_shape
+        img = img[:, :, ::-1]  # bgr2rgb
         resized_image = cv2.resize(
             img, (imgW, imgH), interpolation=cv2.INTER_LINEAR)
         resized_image = resized_image.astype('float32')
diff --git a/tools/infer_rec.py b/tools/infer_rec.py
index a08fa25b..182694e6 100755
--- a/tools/infer_rec.py
+++ b/tools/infer_rec.py
@@ -131,7 +131,6 @@ def main():
             if config['Architecture']['algorithm'] == "SAR":
                 valid_ratio = np.expand_dims(batch[-1], axis=0)
                 img_metas = [paddle.to_tensor(valid_ratio)]
-
             images = np.expand_dims(batch[0], axis=0)
             images = paddle.to_tensor(images)
             if config['Architecture']['algorithm'] == "SRN":
-- 
GitLab