This repository contains materials from the capstone project for the MIT ADSP, 2023
Facial Image Emotion Detection Using Neural Nets (MIT ADSP Capstone Project)
- Built and trained CNNs and transfer learning models (VGG16, ResNetV2, EfficientNet) to classify facial expressions into four emotion categories
- Applied data augmentation and preprocessing techniques to enhance model robustness; automated training and evaluation pipelines in TensorFlow/Keras
- Achieved an F1-score of 0.85 with a fine-tuned VGG model; confusion between 'sad' and 'neutral' reflected labeling ambiguity in the dataset
- Leveraged A100 GPU for training, enabling extensive model experimentation and hyperparameter tuning