Skip to content

Midoriya-w/anomaly-detection-system

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

5 Commits
 
 
 
 
 
 
 
 

Repository files navigation

🔍 Visual Anomaly Detection System

📌 Overview

This project presents an AI-based anomaly detection system for identifying defects in images without requiring labeled defect data. The system uses deep learning techniques to learn normal patterns and detect deviations.

🚀 Features

  • Detects defects without labeled datasets
  • Generates anomaly heatmaps
  • Provides anomaly score
  • Real-time image analysis
  • Simple UI using Streamlit

🧠 Model Used

  • Pre-trained CNN (Feature Extraction)
  • PaDiM (Patch Distribution Modeling) for anomaly detection

⚙️ Tech Stack

  • Python
  • PyTorch
  • OpenCV
  • Anomalib
  • Streamlit

🔄 How It Works

  1. User uploads an image
  2. Image is preprocessed
  3. CNN extracts features
  4. PaDiM models normal feature distribution
  5. System computes anomaly score
  6. Heatmap highlights defective regions

📸 Output Example

▶️ Run the Project

bash pip install -r requirements.txt streamlit run app.py

📂 Project Structure

project ├── app.py ├── model ├── utils ├── images ├── requirements.txt └── README.md

About

AI-based anomaly detection system for industrial defect detection using CNN feature extraction and PaDiM algorithm.

Resources

Stars

Watchers

Forks

Packages

 
 
 

Contributors