silversurfer343/ComputerVisionProject

Wall Crack Detection Application

This application uses both traditional computer vision and deep learning approaches to detect and analyze cracks in wall images. It provides a comprehensive analysis using multiple methods including Harris Corner Detection, Hough Transform, Region-based Segmentation, SLIC, and GMM.

Features

• —Traditional CV Approach:
• —Harris Corner Detection
• —Hough Transform
• —Region-based Segmentation
• —SLIC Segmentation
• —GMM-based Segmentation

• —Deep Learning Approach:
• —DeepLabV3 Model
• —Multiple refinement methods
• —Ensemble predictions

Usage

1. 1.Upload a wall image (JPG/PNG/JPEG)
2. 2.Choose between Traditional CV or Deep Learning approach
3. 3.View detailed analysis and recommendations

Model

The application uses a pre-trained DeepLabV3 model for crack detection.

Requirements

All dependencies are listed in requirements.txt.