N3tron commited on
Commit
ba6a0f5
·
verified ·
1 Parent(s): 0f1d62a

Update video_app.py

Browse files
Files changed (1) hide show
  1. video_app.py +2 -10
video_app.py CHANGED
@@ -15,8 +15,7 @@ import matplotlib.pyplot as plt
15
  import utils
16
  from utils import app
17
 
18
- # Function to perform face recognition in the video
19
- def face_recognition_in_video(video_path, names, embeddings):
20
  # Open the video file
21
  cap = cv2.VideoCapture(video_path)
22
 
@@ -27,13 +26,8 @@ def face_recognition_in_video(video_path, names, embeddings):
27
 
28
  # Define the codec and create VideoWriter object
29
  fourcc = cv2.VideoWriter_fourcc(*'mp4v')
30
- output_path = "/tmp/output_video.mp4"
31
  out = cv2.VideoWriter(output_path, fourcc, fps, (width, height))
32
 
33
- # Initialize FaceAnalysis app
34
- app = FaceAnalysis(name='buffalo_l')
35
- app.prepare(ctx_id=0, det_size=(640, 640))
36
-
37
  # Process each frame of the video
38
  while cap.isOpened():
39
  ret, frame = cap.read()
@@ -49,7 +43,7 @@ def face_recognition_in_video(video_path, names, embeddings):
49
  detected_embedding = face.normed_embedding
50
 
51
  # Calculate similarity scores with known embeddings
52
- scores = np.dot(detected_embedding, np.array(embeddings).T)
53
  scores = np.clip(scores, 0., 1.)
54
 
55
  # Find the index with the highest score
@@ -75,6 +69,4 @@ def face_recognition_in_video(video_path, names, embeddings):
75
  # Release everything if job is finished
76
  cap.release()
77
  out.release()
78
-
79
  return output_path
80
-
 
15
  import utils
16
  from utils import app
17
 
18
+ def face_recognition_in_video(video_path, output_path, names, embeddings):
 
19
  # Open the video file
20
  cap = cv2.VideoCapture(video_path)
21
 
 
26
 
27
  # Define the codec and create VideoWriter object
28
  fourcc = cv2.VideoWriter_fourcc(*'mp4v')
 
29
  out = cv2.VideoWriter(output_path, fourcc, fps, (width, height))
30
 
 
 
 
 
31
  # Process each frame of the video
32
  while cap.isOpened():
33
  ret, frame = cap.read()
 
43
  detected_embedding = face.normed_embedding
44
 
45
  # Calculate similarity scores with known embeddings
46
+ scores = np.dot(detected_embedding, embeddings.T)
47
  scores = np.clip(scores, 0., 1.)
48
 
49
  # Find the index with the highest score
 
69
  # Release everything if job is finished
70
  cap.release()
71
  out.release()
 
72
  return output_path