Collaborate
← Back to projects

May 1, 2020

Live Facial Expression Detection Using Deep CNN (Keras/TensorFlow)

Real-time emotion recognition system trained on grayscale facial images and deployed to an interactive interface.

Delivered end-to-end model training, export, and live inference workflow for expression classification.

TensorFlowKerasOpenCVCNN

Overview

Built and trained a custom convolutional neural network to classify facial expressions into seven emotion categories.

Focus

  • CNN training pipeline on 48x48 grayscale images
  • Face detection and bounding-box extraction with OpenCV
  • Real-time inference for video/image inputs

Project Write-up: https://shamiulshifat.wordpress.com/2020/05/20/project-facial-expressions-detection-using-deep-convolutional-network-using-keras-tensorflow/

Certificate: https://coursera.org/share/5e66c852bf81e730899a9b9c74ea1992