---

Violence Detection

Overview

Violence Detection is a real-time object detection system designed to identify violent elements in images and videos. Using Ultralytics YOLOv11 and a custom dataset managed via Roboflow, the model can accurately detect instances of:

• Blood
• Weapon
• General Violence

---

Features

• Multi-Class Detection: Detects blood, weapon, and violence in real-time.
• Fast & Accurate: Powered by YOLOv11 for high-performance object detection.
• Custom Dataset: Curated and preprocessed with Roboflow for optimal model training.
• Transfer Learning: Fine-tuned YOLO model trained on domain-specific images.

---

Tools

• Ultralytics YOLOv11 (object detection)
• Roboflow (dataset annotation & preprocessing)

---

Tech Stack

• Python
• PyTorch
• OpenCV
• Streamlit

---

Dataset

• Annotated using Roboflow
• Classes: Blood, Weapon, Violence,NonViolence
• Includes image augmentation and preprocessing steps for better generalization

---

Installation

# Clone the repository
git clone https://github.com/yourusername/violence-detection.git
cd violence-detection

# Install dependencies
pip install -r requirements.txt