filter: discard incomplete or half-body poses with low joint visibility

Author: Ian-I-Chien

fall_core.py

@@ -98,13 +98,22 @@ def _safe_aspect_ratio(self, p):
         w, h = max(x) - min(x), max(y) - min(y)
         return w / h if h else 0
 
-    def is_pose_complete(self, pose, required_joints=(10, 11, 13, 14, 22, 23, 25, 26)):
+    def is_pose_complete(self, pose, required_joints=(11, 14, 23, 26)):
         try:
+            complete = True
+            visible_joints = 0
+            length = len(pose) - (len(pose) % 3)
+
+            for i in range(0, length, 3):
+                x, y, conf = pose[i], pose[i + 1], pose[i + 2]
+                if conf > 0.2:
+                    visible_joints += 1
+
             for idx in required_joints:
-                x, y = pose[idx], pose[idx + 1]
-                if x == 0 or y == 0:
-                    return False
-            return True
+                if pose[idx] == 0 or pose[idx + 1] == 0:
+                    complete = False
+
+            return complete and visible_joints >= 10
         except IndexError:
             return False
 
@@ -170,7 +179,7 @@ def draw_debug_overlay(self, image, results):
 
     def load_model(self, path):
         device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
-        weights = torch.load(path, map_location=device)
+        weights = torch.load(path, map_location=device, weights_only=False)
         model = weights["model"].float().eval()
         return (model.half().to(device) if torch.cuda.is_available() else model), device
 

realtime.py

@@ -1,6 +1,6 @@
 import cv2
 import time
-from fall_detector_multi import FallDetectorMulti
+from fall_core import FallDetectorMulti
 from config import FPS, WINDOW_SIZE, V_THRESH, DY_THRESH, ASPECT_RATIO_THRESH
 
 
@@ -20,7 +20,6 @@ def process_realtime_camera():
         if not ret:
             break
 
-        # ✅ 改成使用 core.py 的 handle_frame()
         _image, prev_time = detector.handle_frame(frame, prev_time)
 
         cv2.imshow("Real-Time Fall Detection", _image)

video.py

@@ -8,7 +8,7 @@ def process_video():
         videos_path = "fall_dataset/ci_videos"
         print("[CI MODE] Only running on CI test videos...")
     else:
-        videos_path = "fall_dataset/videos"
+        videos_path = "fall_dataset/test_videos"
 
     output_dir = "output_videos"
     os.makedirs(output_dir, exist_ok=True)