deepinsight / deepinsight/insightface
Why cv2.COLOR_BGR2RGB should be done in face_align_landmark() function?
- Dominant language
- Python
- Stars
- 29.7k
- Forks
- 6.1k
- PR merge metrics
- No merged PRs in 30d
Description
I'd like to do face alignment via calling this function:
https://github.com/deepinsight/insightface/blob/fc622003d5410a64c96024563d7a093b2a55487c/recognition/_evaluation_/ijb/ijb_evals.py#L91
I'm wondering is it necessary to change image channel by `COLOR_BGR2RGB` here?
https://github.com/deepinsight/insightface/blob/fc622003d5410a64c96024563d7a093b2a55487c/recognition/_evaluation_/ijb/ijb_evals.py#L104
Because after calling `face_align_landmark()` I get blue face. Should I do consider this one for getting embedding feature of arcface? Or I could Ignore this one by running another `cv2.cvtColor(ndimage, cv2.COLOR_RGB2BGR)`?
Input image:
detected faces and aligned faces:
Thanks in advance.
Contributor guide
No contributing guide indexed for this repository
Assessment
This issue has not been assessed yet.