Need some advice for face recognition project

im new to with deep learning and computer vision
i want to build my final year project with pytorch , for face recognition in real time !
both detecting and recognition using neural network !
i need some advice from advanced developers who done similar projects?!
i have read alot , blog posts , youtube videos but i still confused , i have my own dataset
and i have a basic understanding of neural network
using a pre-trained model with transform learning is a good choice for those who use google colab service ?
thanks for advice

  • Face detection, get face bounding box and landmarks. Here’s the SOTA Papers and Codes.
  • (Use landmarks) Aligned face is very beneficial for improve the performance of face recognition.
  • Here’s the Megaface Rank.
  • Insightface, which include RetinaFace for face detection and ArcFace for face recognition, use MxNet instead of PyTorch, but you can find some third-party re-implementation in its README file.
  • Dataset quality is more important than network structure.
  • For faster inference: MobileNet v3, EfficientNet and IdleBlock are worth trying.


  • The main idea is to enhance intraclass compactness and inter-class discrepancy.
  • More focused on intraclass compactness: ShpereFace, CosFace and ArcFace are a series.
  • More focused on inter-class discrepancy: UniformFace is one of the recent works.
1 Like