Open Ealuthwala opened 5 years ago
same issue here when using CNN
Well you can use a photo in which face is detected. You can use face haar_cascade file with opencv to capture images whenever a face is detected and later use those images for encoding.
Just saying sometimes it doesn't detect a face without an example isn't really helpful. There are various reasons a face might be missed but it's impossible to know which without seeing a photo that didn't work.
You can use the jitter
parameter in order to increase the distortion of the photos.
in some cases for some photos there are no faces detected during face encoding. is there a better way to make the face encoring without using a pre-captured photo. maybe use a video are a tool that create an save a face encoding through the real time camera