Brick Extraction
William Nickols
Like this project
Posted Jul 6, 2025
AI project that is ongoing, but phase one was successful in extracting text from photos of university bricks in order to compile a list from images.
Likes
1
Views
8
Brick Extraction
AI-Powered Brick OCR Tool for University Sidewalk Archives
What's it doing?
This in-progress AI project leverages computer vision and OCR to extract names and messages from engraved bricks in sidewalk photos—perfect for universities looking to digitize alumni brick paths.
Segmentation at work
The system uses a modular pipeline of:
Preprocessing (for lighting and distortion correction)
Segmentation (to detect and isolate bricks)
Double pass OCR (to extract readable text from each brick)
The extracted data is exported into a structured CSV file, ideal for building searchable alumni databases or powering interactive archives.
Currently, I have a working segmentation demo and a sample output from the OCR pipeline. Once refined, this tool can save universities hundreds of hours of manual data entry and make their donor recognition systems more accessible.
Example data
Tech Stack: Python, OpenCV, Tesseract, NumPy, Pandas
Status: In progress – currently refining real-world segmentation accuracy and UI for easier use. Adding models to improve segmentation.