WRB Crack Detection

Overview

Developed a building WRB crack detection model using Azure Custom Vision for industrial inspection.

Responsibilities

• Collected and ingested dataset
• Trained and evaluated detection model
• Optimized dataset quality and labeling

Approach

• Supervised learning
• Dataset curation
• Model evaluation

Pipeline

flowchart LR
    A[Image Dataset] --> B[Data Labeling]
    B --> C[Model Training]
    C --> D[Evaluation]
    D --> E[Deployment]

Tech

Azure Custom Vision · Deep Learning

Impact

• Enabled automated inspection workflows
• Reduced manual inspection effort
• Improved defect detection consistency