adding app with CLIP image segmentation

app.py

@@ -52,35 +52,29 @@ def detect_using_clip(image,prompts=[],threshould=0.4):
         prompt = prompt.lower()
         
         model_detections[prompt] = [rescale_bbox(prop.bbox,orig_image_shape=image.shape[:2],model_shape=predicted_image.shape[0]) for prop in props]
-        a = np.expand_dims(predicted_image,axis=-1)
-        print(a.shape,image.shape[:2])
-        predicted_images[prompt]= cv2.resize(np.expand_dims(predicted_image,axis=-1),(h,w),interpolation = cv2.INTER_LINEAR)
-
+        predicted_images[prompt]= predicted_image
     return model_detections , predicted_images
 
-def visualize_images(image,detections,predicted_image,prompt):
+def visualize_images(image,detections,predicted_images,prompt):
     alpha = 0.7
-    H,W = image.shape[:2]
+    # H,W = image.shape[:2]
     prompt = prompt.lower()
-    image_copy = image.copy()
-    mask_image = create_mask(image=image_copy,image_mask=predicted_image)
+    image_resize = cv2.resize(image,(352,352))
+    mask_image = create_mask(image=image_resize,image_mask=predicted_images[prompt])
     
     if prompt not in detections.keys():
         print("prompt not in query ..")
-        return image_copy
-    for bbox in detections[prompt]:
-        cv2.rectangle(image_copy, (int(bbox[1]), int(bbox[0])), (int(bbox[3]), int(bbox[2])), (255, 0, 0), 2)
-        cv2.putText(image_copy,str(prompt),(int(bbox[1]), int(bbox[0])),cv2.FONT_HERSHEY_SIMPLEX, 2, 255)
-    final_image = cv2.addWeighted(image_copy,alpha,mask_image,1-alpha,0)
+        return image_resize
+    final_image = cv2.addWeighted(image_resize,alpha,mask_image,1-alpha,0)
     return final_image
 
 def shot(image, labels_text,selected_categoty):
     prompts = labels_text.split(',')
     prompts = list(map(lambda x: x.strip(),prompts))
-
     model_detections,predicted_images  = detect_using_clip(image,prompts=prompts)
 
-    category_image = visualize_images(image=image,detections=model_detections,predicted_image=predicted_images,prompt=selected_categoty)
+    category_image = visualize_images(image=image,detections=model_detections,predicted_images=predicted_images,prompt=selected_categoty)
+    
     return category_image
 
 iface = gr.Interface(fn=shot,