| --- |
| title: Anomaly Vision Agent |
| emoji: 👁️ |
| colorFrom: indigo |
| colorTo: gray |
| sdk: gradio |
| sdk_version: 5.28.0 |
| app_file: app.py |
| pinned: false |
| --- |
| |
| 🌐 **Live Demo**: [Click here to try the Anomaly Vision Agent](https://huggingface.co/spaces/GeetaAIVisionary/SentinelAISurveillanceAnomalyDetection) |
|
|
| # 🧠 Anomaly Vision Agent |
|
|
| **Anomaly Vision Agent** is an AI-powered surveillance intelligence agent that classifies natural-language descriptions of video surveillance events into meaningful categories like **Normal Activity** or **Suspicious Behavior**. |
|
|
| This project uses Hugging Face Transformers and Gradio to provide a lightweight, human-friendly assistant that **simulates how a field security analyst might evaluate real-world behavior**. |
|
|
| ## 🔍 Example Use Cases |
| - “A person is standing at the emergency exit for 20 minutes” → Suspicious |
| - “A janitor is cleaning the hallway during scheduled maintenance” → Normal |
| - “An unknown bag left near gate 4” → Suspicious |
|
|
| ## 🚀 Tech Stack |
| - Hugging Face Transformers (`distilbert-base-uncased-finetuned-sst-2-english`) |
| - Gradio UI |
| - Python + Torch |
| - Semantic logic override for false positives |
|
|
| ## ✨ Created by Geeta Kudumula to demonstrate intelligent surveillance analysis for platforms like **Vaidio** and other AI vision solutions. |
|
|
