The world's simplest facial recognition api for .NET
This repository is porting https://github.com/ageitgey/face_recognition by C#.
This package supports cross platform, Windows, Linux and MacOSX!!
| Package | OS | x86 | x64 | ARM | ARM64 | Nuget |
|---|---|---|---|---|---|---|
| FaceRecognitionDotNet (CPU) | Windows | - | β | - | - |  |
| Linux | - | β | - | - | |
|
| OSX | - | β | - | - | |
|
| FaceRecognitionDotNet for CUDA 9.2 | Windows | - | β | - | - |  |
| Linux | - | β | - | - | |
|
| OSX | - | - | - | - | |
|
| FaceRecognitionDotNet for CUDA 10.0 | Windows | - | β | - | - |  |
| Linux | - | β | - | - | |
|
| OSX | - | - | - | - | |
|
| FaceRecognitionDotNet for CUDA 10.1 | Windows | - | β | - | - |  |
| Linux | - | β | - | - | |
|
| OSX | - | - | - | - | |
|
| FaceRecognitionDotNet for CUDA 10.2 | Windows | - | β | - | - |  |
| Linux | - | β | - | - | |
|
| OSX | - | - | - | - | |
|
| FaceRecognitionDotNet for CUDA 11.0 | Windows | - | β | - | - |  |
| Linux | - | β | - | - | |
|
| OSX | - | - | - | - | |
|
| FaceRecognitionDotNet for CUDA 11.1 | Windows | - | β | - | - |  |
| Linux | - | β | - | - | |
|
| OSX | - | - | - | - | |
|
| FaceRecognitionDotNet for CUDA 11.2 | Windows | - | β | - | - |  |
| Linux | - | β | - | - | |
|
| OSX | - | - | - | - | |
|
| FaceRecognitionDotNet for Intel MKL | Windows | - | β | - | - |  |
| Linux | - | β | - | - | |
|
| OSX | - | β | - | - | |
|
| FaceRecognitionDotNet for ARM | Windows | - | - | - | - |  |
| Linux | - | - | - | - | |
|
| OSX | - | - | - | - | |
β οΈ FaceRecognitionDotNet for ARM is not tested yet
Support API
| face_recognition API | Corresponding API | Note |
|---|---|---|
| batch_face_locations | BatchFaceLocations | |
| compare_faces | CompareFaces | |
| face_distance | FaceDistance | |
| face_encodings | FaceEncodings | |
| face_landmarks | FaceLandmarks | And support Helen dataset |
| face_locations | FaceLocations | And support to get confidence and use custom face detector |
| load_image_file | LoadImageFile | |
| - | CropFaces | Crop image with specified locations |
| - | EyeBlinkDetect | Detect person is blinking or not Support Large model and Helen dataset |
| - | LoadImage | From memory data or System.Drawing.Bitmap |
| - | PredictAge | Predict human age. Use Adience Benchmark Of Unfiltered Faces For Gender And Age Classification dataset |
| - | PredictEmotion | Predict emotion for human face. Use Corrective re-annotation of FER - CK+ - KDEF |
| - | PredictGender | Predict human gender. Use UTKFace dataset |
| - | PredictProbabilityAge | Predict probability of human age. Use Adience Benchmark Of Unfiltered Faces For Gender And Age Classification dataset |
| - | PredictProbabilityEmotion | Predict probability of emotion from human face. Use Corrective re-annotation of FER - CK+ - KDEF |
| - | PredictProbabilityGender | Predict probability of human gender. Use UTKFace dataset |
| - | PredictHeadPose | Predict human head pose. Use 300W-LP dataset |
β οΈ Warning
You must train dataset by yourself. I will NOT provide pretrained model file due to avoiding license issue. You can check the following examples to train dataset.
- tools/AgeTraining
- tools/EmotionTraining
- tools/EmotionTrainingV2
- tools/GenderTraining
- tools/HeadPoseTraining
- tools/HelenTraining
Demo
Face Recognition
Other Face Functions
| Face Landmark | Age and Gender Classification | Head Pose Estimation | Emotion Estimation |
|---|---|---|---|
 |
 |
 |
 |
Document
FaceRecognitionDotNet support full xml document for Visual Studio. A xml document is written English and Japanese. And you can check online document at FaceRecognitionDotNet API Document
Dependencies Libraries and Products
face_recognition
License: The MIT License
Author: Adam Geitgey
Principal Use: The world's simplest facial recognition api for Python and the command line. Main goal of FaceRecognitionDotNet is what ports face_recognition by C#.
face_recognition_models
License: Creative Commons Zero v1.0 Universal License
Author: Adam Geitgey
Principal Use: Trained models for the face_recognition python library
dlib
License: Boost Software License
Author: Davis E. King
Principal Use: A toolkit for making real world machine learning and data analysis applications in C++.
DlibDotNet
License: The MIT License
Author: Takuya Takeuchi
Principal Use: Use dlib interface via .NET. This library is developed by this owner.
OpenCVSharp
License: The BSD 3-Clause License
Author: shimat
Principal Use: Loading image data by opencv wrapper for example