retinaface
facenet
retinaface | facenet | |
---|---|---|
2 | 5 | |
958 | 13,517 | |
- | - | |
7.7 | 0.0 | |
3 days ago | 10 months ago | |
Python | Python | |
MIT License | MIT License |
Stars - the number of stars that a project has on GitHub. Growth - month over month growth in stars.
Activity is a relative number indicating how actively a project is being developed. Recent commits have higher weight than older ones.
For example, an activity of 9.0 indicates that a project is amongst the top 10% of the most actively developed projects that we are tracking.
retinaface
-
what is the best and most optimized model for face detection/face alignment. best for cuda
I tried this implementation https://github.com/serengil/retinaface
-
Using Edge Biometrics For Better AI Security System Development
For face detection, we used the RetinaFace model with a MobileNet backbone from the InsightFace project. This model outputs four coordinates for each detected face on an image as well as 5 facial landmarks. The fact that images captured at different angles or with different optics can change the proportions of the face due to distortion. This may cause the model to struggle identifying the person.
facenet
-
CompreFace - Free and open-source self-hosted face recognition system from Exadel
As for me, openface is already outdated - the latest release was in 2016. If you look for a library, the easiest to use is ageitgey/face_recognition. The more accurate libraries are davidsandberg/facenet and deepinsight/insightface.
-
Facial recognition using cluster
ML training is practically impossible on micro-controllers. Inferencing on the other hand is quite doable, especially if aided by a [TPU coprocessor](https://coral.ai/products/accelerator/). Supposedly with the TPU you can do some quantization-aware training, but I haven't tried this. I am working on a security system that does facial recognition to recognize me and some friends and considers anyone else as an intruder. How I am doing this is by retraining [Facenet](https://github.com/davidsandberg/facenet) with my facial embeddings. Use something like Haar Cascade in OpenCV to get the bounding box for a face and put it through the model to extract face embeddings. You can then save these embeddings as a sort of databases for the faces you want it to recognize during the inferencing phase. After that you can impose something like a SVM classifier to say who in your face database it is. One thing I will note is that the problem is even easier if you are only concerned with one face - in which case it is technically face identification - not recognition. If that is the case, you only need to do a difference calculation between the embeddings you saved during training and the result output from inferencing. If you do end up using the TPU, you can connect to it over USB from inside a container (I only know how to do this in Docker though) too. Hope this was helpful. I am actually looking to use a k8s cluster eventually too as a sort of smart hub for my security system and other devices so I can handle much more traffic (not sure if this is overkill or not on the pi 4s).
-
Man with foot up on desk in Pelosi's office at Capitol arrested
He might just be a solid techie because the scripts are freely available on github. https://github.com/davidsandberg/facenet
What are some alternatives?
yolov8-face - yolov8 face detection with landmark
insightface - State-of-the-art 2D and 3D Face Analysis Project
NeMo - A scalable generative AI framework built for researchers and developers working on Large Language Models, Multimodal, and Speech AI (Automatic Speech Recognition and Text-to-Speech)
deepface - A Lightweight Face Recognition and Facial Attribute Analysis (Age, Gender, Emotion and Race) Library for Python
Face Recognition - The world's simplest facial recognition api for Python and the command line
ECAPA-TDNN - Unofficial reimplementation of ECAPA-TDNN for speaker recognition (EER=0.86 for Vox1_O when train only in Vox2)
CompreFace - Leading free and open-source face recognition system
tiny - Tiny Face Detector, CVPR 2017
DeepStack - The World's Leading Cross Platform AI Engine for Edge Devices
face-alignment - :fire: 2D and 3D Face alignment library build using pytorch
anime-face-detector - Anime Face Detector using mmdet and mmpose