Open my56 opened 4 years ago
Did you check the demos? Literally just do face detection, get the face rectangle, and then grab the face (.getImageData or whatever from a canvas using that rectangle) and pass it to the compute function...very straightforward in the demo code
In gitHub repo(https://github.com/justadudewhohacks/face-api.js) there is a below code to detect or compute face descriptor: const descriptor = await faceapi.computeFaceDescriptor(alignedFaceImage)
My question is HOW to get alignedFaceImage?? Please provide any resources for same to compute face descriptors of two images and match images.