how to use custom face recognition model?

[darkholds]

No description provided.

[takuya-takeuchi]

@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.

[darkholds]

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.

[takuya-takeuchi]

@darkholds
You can use overload to pass custom model files.
https://github.com/takuya-takeuchi/FaceRecognitionDotNet/wiki/Quickstart#1-initialize

[darkholds]

Thank you for that, also can the default recognizer work with custom face detector model,? Or you need to train another recognizer?

[takuya-takeuchi]

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.

[Gaara1996]

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

[takuya-takeuchi]

@Gaara1996
Please do not report issue not related this thread.
I created new issue and please refere this. #190