Update app.py

app.py

@@ -51,8 +51,10 @@ def find_similarity(base64_image, text_input):
 def segment_image(input_image, text_input):
     image_bytes = base64.b64decode(input_image)
     image = Image.open(BytesIO(image_bytes))
-    
-    image = cv2.cvtColor(np.array(image), cv2.COLOR_BGR2RGB)
+
+    # Convert the image to a numpy array
+    image = np.array(image)
+
     mask_generator = SamAutomaticMaskGenerator(sam)
     masks = mask_generator.generate(image)
 
@@ -66,7 +68,7 @@ def segment_image(input_image, text_input):
         cropped_region = segmented_region[y:y+h, x:x+w]
 
         # Convert to base64 image
-        _, buffer = cv2.imencode(".png", cv2.cvtColor(cropped_region, cv2.COLOR_BGR2RGB))
+        _, buffer = cv2.imencode(".png", cropped_region)
         segmented_image_base64 = base64.b64encode(buffer).decode()
 
         # Calculate similarity for the segmented image
@@ -81,6 +83,42 @@ def segment_image(input_image, text_input):
     # Return the segmented images in descending order of similarity
     return segmented_regions
 
+
+
+
+# def segment_image(input_image, text_input):
+#     image_bytes = base64.b64decode(input_image)
+#     image = Image.open(BytesIO(image_bytes))
+    
+#     image = cv2.cvtColor(np.array(image), cv2.COLOR_BGR2RGB)
+#     mask_generator = SamAutomaticMaskGenerator(sam)
+#     masks = mask_generator.generate(image)
+
+#     segmented_regions = []  # List to store segmented regions with similarity scores
+
+#     for i, mask_dict in enumerate(masks):
+#         mask_data = (mask_dict['segmentation'] * 255).astype(np.uint8)
+#         segmented_region = cv2.bitwise_and(image, image, mask=mask_data)
+
+#         x, y, w, h = map(int, mask_dict['bbox'])
+#         cropped_region = segmented_region[y:y+h, x:x+w]
+
+#         # Convert to base64 image
+#         _, buffer = cv2.imencode(".png", cv2.cvtColor(cropped_region, cv2.COLOR_BGR2RGB))
+#         segmented_image_base64 = base64.b64encode(buffer).decode()
+
+#         # Calculate similarity for the segmented image
+#         similarity = find_similarity(segmented_image_base64, text_input)
+
+#         # Append the segmented image and its similarity score
+#         segmented_regions.append({"image": segmented_image_base64, "similarity": similarity})
+
+#     # Sort the segmented images by similarity in descending order
+#     segmented_regions.sort(key=lambda x: x["similarity"], reverse=True)
+
+#     # Return the segmented images in descending order of similarity
+#     return segmented_regions
+
 # Create Gradio components
 input_image = gr.Textbox(label="Base64 Image", lines=8)
 text_input = gr.Textbox(label="Text Input")  # Use Textbox with a label