Update app.py

app.py

@@ -28,10 +28,10 @@ yolo_model = project.version(model_version).model
 
 # ========== Fungsi Deteksi Kombinasi ==========
 def detect_combined(image):
-    # Konversi gambar ke format numpy array
+    # Konversi gambar PIL ke format numpy array
     image_np = np.array(image)
 
-    # Resize gambar ke 640x640
+    # Resize gambar ke 640x640 tanpa mengubah warna atau format
     image_resized = cv2.resize(image_np, (640, 640))
 
     # Simpan gambar input yang sudah di-resize ke file sementara
@@ -43,7 +43,7 @@ def detect_combined(image):
         # ========== [1] YOLO: Deteksi Produk Nestlé ==========
         yolo_pred = yolo_model.predict(temp_path, confidence=50, overlap=80).json()
 
-        # Hitung per kelas dan simpan bounding box (format: (x_center, y_center, width, height))
+        # Hitung jumlah objek per kelas dan simpan bounding box
         nestle_class_count = {}
         nestle_boxes = []
         for pred in yolo_pred['predictions']:
@@ -54,16 +54,13 @@ def detect_combined(image):
         total_nestle = sum(nestle_class_count.values())
 
         # ========== [2] OWLv2: Deteksi Kompetitor ==========
-        headers = {
-            "Authorization": "Basic " + OWLV2_API_KEY,
-        }
-        data = {
-            "prompts": OWLV2_PROMPTS,
-            "model": "owlv2"
-        }
+        headers = {"Authorization": "Basic " + OWLV2_API_KEY}
+        data = {"prompts": OWLV2_PROMPTS, "model": "owlv2"}
+        
         with open(temp_path, "rb") as f:
             files = {"image": f}
             response = requests.post("https://api.landing.ai/v1/tools/text-to-object-detection", files=files, data=data, headers=headers)
+        
         result = response.json()
         owlv2_objects = result['data'][0] if 'data' in result else []
 
@@ -72,7 +69,6 @@ def detect_combined(image):
         for obj in owlv2_objects:
             if 'bounding_box' in obj:
                 bbox = obj['bounding_box']  # Format: [x1, y1, x2, y2]
-                # Filter objek yang sudah terdeteksi oleh YOLO (Overlap detection)
                 if not is_overlap(bbox, nestle_boxes):
                     class_name = obj.get('label', 'unknown').strip().lower()
                     competitor_class_count[class_name] = competitor_class_count.get(class_name, 0) + 1
@@ -89,13 +85,14 @@ def detect_combined(image):
         for class_name, count in nestle_class_count.items():
             result_text += f"{class_name}: {count}\n"
         result_text += f"\nTotal Products Nestle: {total_nestle}\n\n"
+
         if competitor_class_count:
             result_text += f"Total Unclassified Products: {total_competitor}\n"
         else:
             result_text += "No Unclassified Products detected\n"
 
         # ========== [4] Visualisasi ==========
-        img = cv2.imread(temp_path)
+        img = image_resized.copy()
 
         # Gambar bounding box untuk produk Nestlé (Hijau)
         for pred in yolo_pred['predictions']:
@@ -104,7 +101,7 @@ def detect_combined(image):
             cv2.putText(img, pred['class'], (int(x - w/2), int(y - h/2 - 10)),
                         cv2.FONT_HERSHEY_SIMPLEX, 0.5, (0, 255, 0), 2)
 
-        # Gambar bounding box untuk kompetitor (Merah)
+        # Gambar bounding box untuk kompetitor (Merah) dengan label 'unclassified' jika sesuai
         for comp in competitor_boxes:
             x1, y1, x2, y2 = comp['box']
             unclassified_classes = ["cans"]
@@ -113,6 +110,7 @@ def detect_combined(image):
             cv2.putText(img, f"{display_name} {comp['confidence']:.2f}",
                         (int(x1), int(y1 - 10)), cv2.FONT_HERSHEY_SIMPLEX, 0.5, (0, 0, 255), 2)
 
+        # Simpan gambar hasil deteksi
         output_path = "/tmp/combined_output.jpg"
         cv2.imwrite(output_path, img)