kshitizb/emotionlens repository overview

⁠🚀 EmotionLens - AI-Powered Facial Expression Recognition

EmotionLens is an advanced Facial Expression Recognition (FER) web application powered by a custom Residual CNN model. This AI-driven system detects emotions from human faces in real-time with high accuracy.

⁠🔹 Features

✅ FastAPI-based backend
✅ Lightweight CNN architecture (No pre-trained models!)
✅ Supports drag-and-drop, paste, and file upload
✅ Predicts emotions with confidence scores
✅ Fully containerized for quick deployment

⁠📌 Usage

⁠Run the Container

docker run -p 8010:8010 kshitizb/emotionlens

Then open: http://localhost:8010/⁠

⁠Pull the Image

docker pull kshitizb/emotionlens

⁠📌 Supported API Endpoints

⁠🔹 Recognized Emotions

😠 Anger
😒 Contempt
🤢 Disgust
😨 Fear
😃 Happy
😐 Neutral
😢 Sad
😲 Surprise

⁠📌 Deployment

⁠Build Locally

docker build -t kshitizb/emotionlens .

⁠Push to Docker Hub

docker tag emotionlens kshitizb/emotionlens
docker push kshitizb/emotionlens

⁠📌 Model Details

• Dataset: Trained on AffectNet
• Architecture: Residual CNN
• Accuracy: 73.96% (on test set)

⁠🔹 Contributors

👨‍💻 Kshitiz Bhargava⁠
👨‍💻 Manvendra Singh
👨‍💻 Krishnaraj Natarajan

📧 Contact: [email protected]⁠

⁠🔹 License

📝 Open-source under MIT License.