DynaDetect AI — Hidden Leaf Vision System

Advanced Zero-Shot Object Detection & Target Registry Portal

DynaDetect AI is a dynamic computer vision application built to mimic the analytical precision of the Sharingan. By integrating state-of-the-art edge AI models with a high-performance backend, it provides real-time, zero-shot object detection capabilities through an immersive, themed web interface.

---

Key Features

• Sharingan Vision Interface: A highly responsive, custom-designed frontend dashboard built with HTML, CSS, and Vanilla JavaScript.
• Zero-Shot Object Detection: Utilizes Grounding DINO to detect and track any object purely based on text prompts—no rigid pre-training categories required.
• Real-Time Analytics: Integrates YOLO (You Only Look Once) for high-speed, standard object detection bounding boxes.
• Target Registry System: A dedicated UI section that logs and catalogs detected objects dynamically.
• High-Speed API: Backend powered by FastAPI, ensuring rapid inference times and smooth communication between the web portal and the AI models.

---

Tech Stack

Frontend:

• HTML5 / CSS3
• Vanilla JavaScript
• Custom Fonts (Orbitron, Rajdhani, Share Tech Mono)

Backend & AI:

• Python
• FastAPI
• YOLO (Ultralytics)
• Grounding DINO
• Uvicorn (ASGI Server)

---

Local Installation & Setup

Since this project relies on heavy AI models (YOLO, Grounding DINO) and a FastAPI backend, the web server needs to run locally to process the video/image feeds.

Prerequisites

• Python 3.8+
• Git
• A webcam (for real-time detection)

1. Clone the Repository

git clone [https://github.com/skjha3439/dynadetect-ai.git](https://github.com/skjha3439/dynadetect-ai.git)
cd dynadetect-ai