Faqe is a face recognition client, built with React. To try the application, the first step is to take a reference image of your face on the reference Tab. Then head up to the Recogntion Tab to perform recognition. You can set a recogntion threshold, which will figure how strict should the system be.
Status icons on the Recogntion Tab provided by Vitaly Gorbachev on Flaticon. Icons used for Tabs are from Good Ware, Chris Veigt and Freepik also from Flaticon.
- For face detection, this project implements a SSD (Single Shot Multibox Detector) based on MobileNetV1, trained on the WIDERFACE dataset and the weights are provided by yeephycho in this repo. This API and the next are both provided by this project.
- For face recognition, a ResNet-34 like architecture is implemented to compute a face descriptor (a feature vector with 128 values) from any given face image, which is used to describe the characteristics of a persons face. The weights have been trained by davisking and the model achieves a prediction accuracy of 99.38% on the LFW (Labeled Faces in the Wild) benchmark for face recognition.
Faqe is a client software, and was only created for an educational purposes. The reference image taken by the app is not being sent to a server or to a database for processing or other uses, and only a 128-dimensionnal space embedding (representation) of your face is stored on your local machine for face recognition. The face recognition systems works the same way, all on your computer, and make sure to contact me if there are any questions on the app.