Human Emotion Detection

This project explores multiple convolutional architectures for detecting human emotions from facial expressions.

---

Example Prediction After Training ResNet34 model for 60 epochs

CNN (2 Conv2D)

• 20 Epochs

Accuracy over 20 epochs	Loss over 20 epochs

CUTMIX

• 3 epochs

Accuracy: 3 epochs	Loss: 3 epochs

• 20 epochs

Accuracy: 20 epochs	Loss: 20 epochs

ResNet

• ResNet34 with 60 epochs

Accuracy: 60 epochs	Loss: 60 epochs

Transfer Learning With EfficientNet

ViT