From f18ddffabd065f5ba176c2fb6e21cb45d2c6fe8a Mon Sep 17 00:00:00 2001
From: zhoujun <572459439@qq.com>
Date: Wed, 16 Nov 2022 15:54:36 +0800
Subject: [PATCH] update inference of PP-OCRv2 to paddlecv (#5587)

* add install of paddleclas

* update inference with paddlecv
---
 modelcenter/PP-OCRv2/APP/app.py       | 26 ++++++++++----------------
 modelcenter/PP-OCRv3/APP/app.py       | 26 ++++++++++----------------
 modelcenter/PP-StructureV2/APP/app.py |  9 ++++-----
 3 files changed, 24 insertions(+), 37 deletions(-)

diff --git a/modelcenter/PP-OCRv2/APP/app.py b/modelcenter/PP-OCRv2/APP/app.py
index 28dcc843..7de73c1d 100644
--- a/modelcenter/PP-OCRv2/APP/app.py
+++ b/modelcenter/PP-OCRv2/APP/app.py
@@ -3,10 +3,8 @@ import base64
 from io import BytesIO
 from PIL import Image
 
-import paddleocr
-from paddleocr import PaddleOCR, draw_ocr
-
-ocr = PaddleOCR(ocr_version='PP-OCRv2', use_angle_cls=True, lang="ch")
+from paddlecv import PaddleCV
+ocr = PaddleCV(task_name="PP-OCRv2")
 
 
 def image_to_base64(image):
@@ -20,20 +18,16 @@ def image_to_base64(image):
 
 # UGC: Define the inference fn() for your models
 def model_inference(image):
-    result = ocr.ocr(image, cls=True)
-
-    # 显示结果
-    if paddleocr.__version__ >= "2.6.0.2":
-        result = result[0]
-    boxes = [line[0] for line in result]
-    txts = [line[1][0] for line in result]
-    scores = [line[1][1] for line in result]
-    im_show = draw_ocr(image, boxes, txts=None, scores=None)
-    im_show = Image.fromarray(im_show)
+    result = ocr(image)[0]
 
+    im_show = Image.open('output/tmp.jpg')
     res = []
-    for i in range(len(boxes)):
-        res.append(dict(boxes=boxes[i], txt=txts[i], score=scores[i]))
+    for i in range(len(result['dt_polys'])):
+        res.append(
+            dict(
+                boxes=result['dt_polys'][i],
+                txt=result['rec_text'][i],
+                score=result['rec_score'][i]))
     json_out = {"base64": image_to_base64(im_show), "result": res}
     return im_show, json_out
 
diff --git a/modelcenter/PP-OCRv3/APP/app.py b/modelcenter/PP-OCRv3/APP/app.py
index 04b7f8be..9fe9b6ff 100644
--- a/modelcenter/PP-OCRv3/APP/app.py
+++ b/modelcenter/PP-OCRv3/APP/app.py
@@ -3,10 +3,8 @@ import base64
 from io import BytesIO
 from PIL import Image
 
-import paddleocr
-from paddleocr import PaddleOCR, draw_ocr
-
-ocr = PaddleOCR(use_angle_cls=True, lang="ch")
+from paddlecv import PaddleCV
+ocr = PaddleCV(task_name="PP-OCRv3")
 
 
 def image_to_base64(image):
@@ -20,20 +18,16 @@ def image_to_base64(image):
 
 # UGC: Define the inference fn() for your models
 def model_inference(image):
-    result = ocr.ocr(image, cls=True)
-
-    # 显示结果
-    if paddleocr.__version__ >= "2.6.0.2":
-        result = result[0]
-    boxes = [line[0] for line in result]
-    txts = [line[1][0] for line in result]
-    scores = [line[1][1] for line in result]
-    im_show = draw_ocr(image, boxes, txts=None, scores=None)
-    im_show = Image.fromarray(im_show)
+    result = ocr(image)[0]
 
+    im_show = Image.open('output/tmp.jpg')
     res = []
-    for i in range(len(boxes)):
-        res.append(dict(boxes=boxes[i], txt=txts[i], score=scores[i]))
+    for i in range(len(result['dt_polys'])):
+        res.append(
+            dict(
+                boxes=result['dt_polys'][i],
+                txt=result['rec_text'][i],
+                score=result['rec_score'][i]))
     json_out = {"base64": image_to_base64(im_show), "result": res}
     return im_show, json_out
 
diff --git a/modelcenter/PP-StructureV2/APP/app.py b/modelcenter/PP-StructureV2/APP/app.py
index 3637161a..c88998f8 100644
--- a/modelcenter/PP-StructureV2/APP/app.py
+++ b/modelcenter/PP-StructureV2/APP/app.py
@@ -3,9 +3,8 @@ import base64
 from io import BytesIO
 from PIL import Image
 
-from paddleocr import PPStructure
-
-table_engine = PPStructure(layout=False, show_log=True)
+from paddlecv import PaddleCV
+table_engine = PaddleCV(task_name="PP-StructureV2-table")
 
 
 def image_to_base64(image):
@@ -19,8 +18,8 @@ def image_to_base64(image):
 
 # UGC: Define the inference fn() for your models
 def model_inference(image):
-    result = table_engine(image)
-    res = result[0]['res']['html']
+    result = table_engine(image)[0]
+    res = result['html']
     json_out = {"result": res}
     return res, json_out
 
-- 
GitLab