Skip to content

Latest commit

 

History

History
24 lines (16 loc) · 674 Bytes

README.md

File metadata and controls

24 lines (16 loc) · 674 Bytes
Raw

real-time emotion recognition

A real-time emotion recognition project which can run on CPU.

Requirements

  • python 3.6
  • opencv 3.4.1
  • tensorflow 1.5
  • keras 2.1.5

Running

Using pc camera to do real-time emotion recognition.

python emotion_recognition_cam.py

Pipline

  1. Using MTCNN model to detecte face and its landmarks.
  2. Using RAF-DB emotion dateset(aligned) to train a mobilenet model (test acc=86.05%).