Problem "ValueError: operands could not be broadcast together with shapes ( , ) (128, )" Solution

[MrLangZe]

• face_recognition version:
• Python version: 3.8.18
• Operating System: Windows 11

Description

When I tried to run the /examples/facerec_from_webcam_faster.py, I found the error which is "ValueError: operands could not be broadcast together with shapes ( , ) (128, )".
There may be three possible reasons to cause the problem which I will show below.

What I Did

1.Check your image file encoding index.Get your image encoding by this way:
your_image = face_recognition.load_image_file("../your.jpg")
your_face_encoding = face_recognition.face_encodings(your_image)[0]

2.The parameter "face_location" in method "face_encodings" may not need when you first run the demo.So you can change the code like this:
face_locations = face_recognition.face_locations(rgb_small_frame)
face_encodings = face_recognition.face_encodings(rgb_small_frame)

3.The BGR image can be converted to RGB image by using cv2.cvtColor like this:
rgb_small_frame = cv2.cvtColor(small_frame, cv2.COLOR_BGR2RGB)