Extract information from videos/images using OCR + Llama 3.1 8B model with real-time anomaly detection.
Complete pipeline: Video/Image Input → OCR Processing → LLM Analysis → Anomaly Detection → Results Visualization
- OCR Analysis: EasyOCR/Tesseract text extraction
- LLM Intelligence: Llama 3.1 8B content analysis
- Anomaly Detection: ResNet-18 + IsolationForest two-stage model
- Interactive UI: Streamlit with real-time visualization
- Production Ready: Docker, modular architecture, comprehensive tests
- Llama 3.1 8B model
docker build -t journeyspotter .
GROQ_API_KEY=your_groq_api_key_here docker run -p 8000:8000 -e GROQ_API_KEY journeyspotter