Update app.py

app.py

@@ -195,7 +195,23 @@ async def predict_single_dog(image):
     
 #     return dogs
 
-async def detect_multiple_dogs(image, conf_threshold=0.25, iou_threshold=0.6):
+
+def filter_detections(dogs, image_size):
+    filtered_dogs = []
+    image_area = image_size[0] * image_size[1]
+    
+    for dog in dogs:
+        _, _, box = dog
+        dog_area = (box[2] - box[0]) * (box[3] - box[1])
+        area_ratio = dog_area / image_area
+        
+        if 0.01 < area_ratio < 0.9:  # 過濾掉太小或太大的檢測框
+            filtered_dogs.append(dog)
+    
+    return filtered_dogs
+    
+
+async def detect_multiple_dogs(image, conf_threshold=0.2, iou_threshold=0.4):
     results = model_yolo(image, conf=conf_threshold, iou=iou_threshold)[0]
     dogs = []
     boxes = []
@@ -217,32 +233,11 @@ async def detect_multiple_dogs(image, conf_threshold=0.25, iou_threshold=0.6):
             
             if is_valid_dog(cropped_image):
                 dogs.append((cropped_image, confidence, [x1, y1, x2, y2]))
+
+    dogs = filter_detections(dogs, (image.width, image.height))
     
     return dogs
 
-def is_valid_dog(image):
-    img_array = np.array(image)
-    
-    # 1. 簡單的紋理檢測
-    gray = np.mean(img_array, axis=2)
-    texture = np.std(gray)
-    
-    # 2. 顏色分布檢測
-    img_rgb = img_array.reshape(-1, 3)
-    kmeans = KMeans(n_clusters=3, n_init=10)
-    kmeans.fit(img_rgb)
-    colors = kmeans.cluster_centers_
-    color_variety = np.std(colors)
-    
-    # 3. 形狀檢測（檢查長寬比和面積）
-    aspect_ratio = image.width / image.height
-    area_ratio = (image.width * image.height) / (image.width * image.height)
-    
-    # 放寬判斷條件
-    if (texture > 5 and color_variety > 20 and 
-        0.3 < aspect_ratio < 3 and area_ratio > 0.05):
-        return True
-    return False
 
 def merge_overlapping_boxes(boxes, overlap_threshold):
     merged = []