🎥 Video Emotion SaaS

A production-grade Software-as-a-Service (SaaS) platform that serves your Video Emotion Recognition AI model through an intuitive web interface and secure REST API.
Users can upload videos, analyze emotions and sentiments, and access insights both visually and programmatically.

---

🧠 About

Video Emotion SaaS transforms your trained deep learning model into a full cloud-based service.
It combines a modern web UI, scalable backend, and API integration to make emotion recognition accessible, interactive, and developer-friendly.

Built with AWS SageMaker, Amazon S3, FastAPI, and React, it represents the production layer of the Video Emotion Project.

---

📸 UI Snapshots

Dashboard & Inference View

Registration Page

---

🏗️ Architecture Overview

[Frontend UI]  →  [API Gateway / FastAPI Backend]  →  [AWS SageMaker Endpoint]
         ↘                   ↕
         [Auth Service]    [Task Queue / Worker]
                            ↕
                     [S3 Storage / Database]

Core Components

Layer	Description
Frontend	Built with React, it provides video uploads, visual analytics, and real-time emotion timelines
Backend (FastAPI)	Exposes REST APIs, manages authentication, and orchestrates inference requests
Worker	Handles background processing and asynchronous tasks
Model Endpoint	Hosted on AWS SageMaker, providing emotion recognition via API calls
Storage	Videos and results stored securely on AWS S3
Database	Manages users, job tracking, and analytics summaries
Auth	JWT and OAuth2-based secure authentication

---

✨ Features

• 🎞️ Upload and process videos through a clean, modern UI
• 🔐 User authentication and secure API key management
• ⚙️ Asynchronous video processing via task queues
• 📊 Emotion and sentiment visualization dashboard
• 🧾 REST API for developers and integrations
• ☁️ AWS-based scalable architecture
• 🧠 Powered by deep learning (CNN + LSTM model) via SageMaker

---

🧰 Technologies Used

Category	Technology	Description
Frontend	React / TypeScript	User dashboard and upload interface
Backend	FastAPI / Python	REST API and inference orchestration
AI Model	AWS SageMaker	Deployed emotion detection model
Storage	AWS S3	Persistent storage for videos and analysis data
Database	PostgreSQL / DynamoDB	Job management and analytics
Queue / Worker	Celery / Redis	Background task processing
Auth	JWT / OAuth2	Secure authentication and user sessions
Deployment	Docker / ECS / CI/CD	Scalable deployment workflow
Monitoring	CloudWatch / Grafana	System monitoring and alerts

---

🚀 Setup & Usage

1️⃣ Prerequisites

• AWS account with S3 and SageMaker access
• Python 3.8+
• Docker (optional)
• .env configuration file

2️⃣ Clone & Install

git clone https://github.com/danielgavrila2/Video-Emotion-SaaS.git
cd Video-Emotion-SaaS
pip install -r requirements.txt

3️⃣ Configure Environment

Create a .env file with:

MODEL_ENDPOINT_URL=https://<your-sagemaker-endpoint>
S3_BUCKET=your-bucket-name
DATABASE_URL=postgresql://user:pass@host:port/db
SECRET_KEY=supersecretkey
AWS_REGION=us-east-1

4️⃣ Run Locally

uvicorn app.main:app --reload --port 8000

Frontend (if separate):

cd frontend && npm install && npm start

---

🔗 API Reference

Method	Endpoint	Description
POST	/api/v1/predict	Upload a video for emotion inference
GET	/api/v1/jobs/{job_id}	Retrieve prediction results
POST	/api/v1/register	Create new user account
POST	/api/v1/login	Authenticate and obtain JWT token
GET	/api/v1/users/me	Get user profile information

Example Request

curl -X POST "http://localhost:8000/api/v1/predict"   -H "Authorization: Bearer <token>"   -F "video=@/path/to/video.mp4"

Response:

{
  "job_id": "12345",
  "status": "processing"
}

---

☁️ Deployment

Deploy using Docker or AWS ECS:

docker-compose up --build

Example Docker Compose snippet:

version: "3"
services:
  backend:
    build: ./backend
    ports:
      - "8000:8000"
    env_file: .env
  frontend:
    build: ./frontend
    ports:
      - "3000:3000"

For production, integrate with:

• AWS ECS or EKS for scalability
• AWS CloudWatch for monitoring
• GitHub Actions for CI/CD automation

---

🔒 Security

• Encrypted communication (HTTPS)
• JWT-based authentication
• Role-based access (User/Admin)
• File validation for uploads
• Quota & rate limiting per user
• Logging and audit trails

---

📈 Future Enhancements

• Real-time inference feedback
• Advanced analytics dashboard
• Multi-model support (emotion + sentiment + tone)
• Subscription-based billing integration
• Model version management

---

🤝 Contributing

Contributions are welcome!

1. Fork this repo
2. Create a branch: feature/your-feature
3. Commit and push your changes
4. Open a Pull Request

---

📜 License

Released under the MIT License.
© 2025 Daniel Gavrila. All rights reserved.

---

🧩 Repository Links

• Model Training Project: Video Emotion Project
• SaaS Deployment: Video Emotion SaaS