Update app.py

app.py

@@ -16,37 +16,33 @@ img_height, img_width = 150, 150  # Input size for the model
 def predict_labels_and_probabilities(image_path):
     # Load the image using OpenCV
     img = cv2.imdecode(np.frombuffer(image_path.read(), np.uint8), cv2.IMREAD_COLOR)
+    img_rgb = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
     
-    # Assuming the input image might contain multiple ingredients, we will process it in patches.
-    # For simplicity, let's divide the image into 4 patches and classify each one
-    h, w, _ = img.shape
-    patches = [
-        img[0:h//2, 0:w//2],  # Top-left
-        img[0:h//2, w//2:w],  # Top-right
-        img[h//2:h, 0:w//2],  # Bottom-left
-        img[h//2:h, w//2:w],  # Bottom-right
-    ]
-    
+    # Scan the image in a grid-like fashion
+    step_size = 100  # Step size for the grid
     predictions = []
     
-    for patch in patches:
-        patch_resized = cv2.resize(patch, (img_height, img_width))
-        patch_array = image.img_to_array(patch_resized)
-        patch_array = np.expand_dims(patch_array, axis=0)
-        patch_array /= 255.  # Scale pixel values
-        
-        preds = model.predict(patch_array)
-        class_idx = np.argmax(preds[0])
-        
-        # Map class indices to class names
-        class_labels = {
-            0: 'Bell Pepper', 1: 'Carrot', 2: 'Garlic', 3: 'Ginger',
-            4: 'Jalapeno', 5: 'Onion', 6: 'Potato', 7: 'Sweetpotato', 8: 'Tomato'
-        }
-        predicted_class = class_labels[class_idx]
-        probability = preds[0][class_idx]
-        
-        predictions.append((predicted_class, probability))
+    for y in range(0, img.shape[0] - img_height, step_size):
+        for x in range(0, img.shape[1] - img_width, step_size):
+            patch = img_rgb[y:y+img_height, x:x+img_width]
+            patch_resized = cv2.resize(patch, (img_height, img_width))
+            patch_array = image.img_to_array(patch_resized)
+            patch_array = np.expand_dims(patch_array, axis=0)
+            patch_array /= 255.  # Scale pixel values
+            
+            preds = model.predict(patch_array)
+            class_idx = np.argmax(preds[0])
+            
+            # Map class indices to class names
+            class_labels = {
+                0: 'Bell Pepper', 1: 'Carrot', 2: 'Garlic', 3: 'Ginger',
+                4: 'Jalapeno', 5: 'Onion', 6: 'Potato', 7: 'Sweetpotato', 8: 'Tomato'
+            }
+            predicted_class = class_labels[class_idx]
+            probability = preds[0][class_idx]
+            
+            if probability > 0.5:  # Threshold to filter out low confidence predictions
+                predictions.append((predicted_class, probability, x, y, x+img_width, y+img_height))
     
     return predictions
 
@@ -58,12 +54,19 @@ uploaded_file = st.file_uploader("Choose an ingredients image...", type="jpg")
 if uploaded_file is not None:
     # Display the uploaded image
     st.image(uploaded_file, caption='Uploaded Ingredient Image.', use_column_width=True)
-
+    
     # Perform the prediction and display the results
     predictions = predict_labels_and_probabilities(uploaded_file)
-    st.write("Predictions for different patches of the image:")
+    st.write("Predictions for detected ingredients in the image:")
+    
+    for i, (label, probability, x1, y1, x2, y2) in enumerate(predictions):
+        st.write(f"Ingredient {i+1}:")
+        st.write(f"Prediction: {label}")
+        st.write(f"Probability: {probability}")
+        st.write(f"Location: ({x1}, {y1}) to ({x2}, {y2})")
+        
+        # Draw rectangle around detected ingredients
+        cv2.rectangle(img_rgb, (x1, y1), (x2, y2), (255, 0, 0), 2)
     
-    for i, (label, probability) in enumerate(predictions):
-        st.write(f"Patch {i+1}:")
-        st.write("Prediction:", label)
-        st.write("Probability:", probability)
+    # Display the image with rectangles
+    st.image(img_rgb, caption='Detected Ingredients', use_column_width=True)