how to use custom face recognition model?
darkholds opened this issue · comments
@darkholds
FRDN does not provide tools to generate model for recognition.
Btt officiai dlib blog page indicates about deep mertric learning.
http://blog.dlib.net/2017/02/high-quality-face-recognition-with-deep.html
You may get useful information from the above link.
hi, for example, i have a model (dat file) generated from dlibdotnet, how can i use it in FRDN? is there any workaround? i want to recognise people with facemask if possible. Thanks.
@darkholds
You can use overload to pass custom model files.
https://github.com/takuya-takeuchi/FaceRecognitionDotNet/wiki/Quickstart#1-initialize
Thank you for that, also can the default recognizer work with custom face detector model,? Or you need to train another recognizer?
You could use custom mode which is generated from same model type. i.e. ShapePredictor.
https://github.com/takuya-takeuchi/DlibDotNet/blob/master/examples/TrainShapePredictor/Program.cs
But I have not tested yet.
Hello
I try to do facial recognition with this example
https://github.com/takuya-takeuchi/FaceRecognitionDotNet/wiki/Quickstart#1-initialize
but I get an error when compiling in visual studio
Argument 2: cannot be converted from 'FaceRecognitionDotNet.Location' a 'System.Collections.Generic.IEnumerable<FaceRecognitionDotNet.Location>'
in this line
IEnumerable encodingA = FaceRecognition.FaceEncodings(imageA, locationsA.First());
I am missing some nuget package
@Gaara1996
Please do not report issue not related this thread.
I created new issue and please refere this. #190