--- a
+++ b/demo-fall-detection-cmd.py
@@ -0,0 +1,47 @@
+import argparse
+from pathlib import Path
+from PIL import Image
+from fall_prediction import Fall_prediction
+import numpy as np
+import json
+import time
+import yaml
+
+class JsonEncoder(json.JSONEncoder):
+    def default(self, obj):
+        if isinstance(obj, np.integer):
+            return int(obj)
+        if isinstance(obj, np.floating):
+            return float(obj)
+        if isinstance(obj, np.ndarray):
+            return obj.tolist()
+
+        return super(JsonEncoder, self).default(obj)
+
+parser = argparse.ArgumentParser()
+parser.add_argument("--image_1", type=Path, help='Path to the First Image', required=True)
+parser.add_argument("--image_2", type=Path, help='Path to the Second Image', required=True)
+parser.add_argument("--image_3", type=Path, help='Path to the Third Image')
+
+p = parser.parse_args()
+img1 = Image.open(p.image_1)
+img2 = Image.open(p.image_2)
+img3 = Image.open(p.image_3) if p.image_3 else None
+
+response = Fall_prediction(img1, img2, img3)
+
+if response:
+
+    print("There is", response['category'])
+    print("Confidence :", response['confidence'])
+    print("Angle : ", response['angle'])
+    print("Keypoint_corr :", response['keypoint_corr'])
+
+    time_str = time.strftime("%Y%m%d-%H%M%S")
+    json_str = json.dumps(response, cls=JsonEncoder)
+
+    with open(f"tmp/{time_str}.yaml","w",) as file:
+        yaml.dump(json.loads(json_str), file)
+else:
+     print("There is no fall detetcion...")
+