atulapra/Emotion-detection

Classifies facial expressions into seven emotion categories (angry, disgusted, fearful, happy, neutral, sad, surprised) using a 4-layer CNN trained on the FER-2013 dataset. The pipeline combines Haar cascade face detection for real-time frame analysis with 48x48 grayscale image normalization before CNN inference, outputting softmax probability scores for each emotion class. Built on TensorFlow 2.0 with Keras API and OpenCV for webcam integration, achieving 63.2% test accuracy.

1,346 stars. No commits in the last 6 months.