Create app.py

app.py

@@ -0,0 +1,51 @@
+import os
+os.system("pip install --upgrade pip")
+os.system("pip install torch torchvision")
+os.system("pip install numpy==1.21.0 mmcv-full openmim requests")
+os.system("mim install mmdet")
+os.system('git clone https://github.com/open-mmlab/mmocr.git')
+os.system('pip install -r mmocr/requirements.txt')
+os.system('pip install -v -e mmocr')
+os.system('export PYTHONPATH=$(pwd)/mmocr:$PYTHONPATH')
+os.system('pip freeze')
+import sys
+import requests
+sys.path.append('./mmocr')
+from gradio import inputs,outputs,Interface
+from PIL import Image
+from mmocr.utils.ocr import MMOCR
+
+# load image examples
+urls = ['https://i0.wp.com/www.kento.com.my/wp-content/uploads/2020/09/ikhsan-sports-shop-3D-stainless-steel-kento.jpg', 'https://automology.s3.ap-southeast-1.amazonaws.com/wp-content/uploads/2019/08/12224246/P.-Mampos.jpg']
+for idx, url in enumerate(urls):
+  print(url)
+  image = Image.open(requests.get(url, stream=True).raw)
+  image.save(f"image_{idx}.png")
+det = ["DB_r18","DB_r50","DRRG","FCE_IC15","MaskRCNN_CTW","MaskRCNN_IC15","MaskRCNN_IC17","PS_CTW","PS_IC15","TextSnake"]
+rec = ["CRNN","SAR","NRTR_1/16-1/8","NRTR_1/8-1/4","RobustScanner","SATRN","SATRN_sm","ABINet","SEG","CRNN_TPS"]
+# [MMOCR(config_dir="mmocr/configs/",det=i,recog=None,device="cpu") for i in det]
+# [MMOCR(config_dir="mmocr/configs/",det=None,recog=i,device="cpu") for i in rec]
+
+def process_image(detection_model_name,recognizer_model_name,image_path):
+    ocr = MMOCR(config_dir="mmocr/configs/",det=detection_model_name,recog=recognizer_model_name,device="cpu")
+    results = ocr.readtext(image_path,output='demo/det_out.jpg')
+    img = Image.open("demo/det_out.jpg")
+    return [img,results[0]["text"]]
+title = "Interactive demo: MMOCR"
+description = "OpenMMLab Text Detection, Recognition and Understanding Toolbox"
+article = "Github repo: https://github.com/open-mmlab/mmocr"
+examples =[["TextSnake","RobustScanner",f"image_{i}.png"] for i in range(len(urls))]
+#css = """.output_image, .input_image {height: 600px !important}"""
+iface = Interface(fn=process_image,
+                     inputs=[
+                        inputs.Dropdown(det,type="value"),
+                        inputs.Dropdown(rec,type="value"),
+                        inputs.Image(type="filepath")
+                     ],
+                     outputs=[outputs.Image(type="pil"),outputs.Textbox()],
+                     title=title,
+                     description=description,
+                     article=article,
+                     examples=examples
+                    )
+iface.launch(debug=False)