refactor: change cv2 to pillow to read image

main.py

@@ -40,29 +40,22 @@ def classify_image(img):
     return confidences
 
 def animal_detect_and_classify(img_path):
-    # Read the image
-    img = cv2.imread(img_path)
+    # Read the image using Pillow
+    img = Image.open(img_path)
 
     # Pass the image through the detection model and get the result
-    detect_results = detection_model(img)
+    detect_results = detection_model(np.array(img))
 
     combined_results = []
-    # print("dss", detect_results[0])
-    # Iterate over the detected objects
+
     # Iterate over detections
     for result in detect_results:
         flag = False
         for box in result.boxes:
             flag = True
-            # print(box)
-            # Crop the RoI
+            # Crop the Region of Interest (RoI)
             x1, y1, x2, y2 = map(int, box.xyxy[0])
-            detect_img = img[y1:y2, x1:x2]
-            # Convert the image to RGB format
-            detect_img = cv2.cvtColor(detect_img, cv2.COLOR_BGR2RGB)
-
-            # Resize the input image to the expected shape (224, 224)
-            detect_img = cv2.resize(detect_img, (224, 224))
+            detect_img = img.crop((x1, y1, x2, y2)).resize((224, 224))
 
             # Convert the image to a numpy array
             inp_array = np.array(detect_img)
@@ -75,6 +68,7 @@ def animal_detect_and_classify(img_path):
 
             # Make predictions using the classification model
             prediction = classification_model.predict(inp_array)
+
             # Map predictions to labels
             threshold = 0.66
             predicted_labels = [labels[np.argmax(pred)] if np.max(pred) >= threshold else "animal" for pred in prediction]
@@ -82,12 +76,10 @@ def animal_detect_and_classify(img_path):
             combined_results.append(((x1, y1, x2, y2), predicted_labels))
         if flag:
             continue
-        y2, x2, _ = img.shape
-        detect_img = img[0:y2, 0:x2]
-        detect_img = cv2.cvtColor(detect_img, cv2.COLOR_BGR2RGB)
-        detect_img = cv2.resize(detect_img, (224, 224))
-        inp_array = np.array(detect_img)
-        inp_array = inp_array.reshape((-1, 224, 224, 3))
+        # If no detections found, consider the whole image
+        x2, y2 = img.size
+        detect_img = img.resize((224, 224))
+        inp_array = np.array(detect_img).reshape((-1, 224, 224, 3))
         inp_array = tf.keras.applications.efficientnet.preprocess_input(inp_array)
         prediction = classification_model.predict(inp_array)
         threshold = 0.66