RhizoScan

• 

  Landing page

• 

  Processing Image processing

• 

  After Image processing

• 

  Image segementation

Inspiration

In plant science, particularly root biology, researchers often struggle with accurately quantifying root traits due to cumbersome software setups, inconsistent image scaling, and reliance on external hardware. We were inspired to create RhizoScan after realizing that many labs, especially in developing regions, lack easy-to-use tools for root trait analysis. Our goal was to create a lightweight, accessible, and precise tool that runs entirely in the browser and works with just a smartphone and a sheet of A4 paper.

What it does

RhizoScan enables researchers to analyze root images easily and accurately. Users take a photo of a plant root placed on A4 paper, and the tool: Automatically detects and segments the root structure. Uses the A4 paper as a real-world reference to calculate: Root length Area Perimeter Circularity Solidity Aspect ratio Outputs measurement data in standard units (e.g., mm). Runs entirely in the browser, with no login, no image upload, and no installations.

How we built it

First, I use chatGPT to generate a comprehensive prompt needed. I transferred the prompts in bolt.new. I tested the generated application. Then I deployed.

Challenges we ran into

Creation of proper image and calibration of result.

Accomplishments that we're proud of

Building a fully functional root analysis tool that works offline and in-browser. Achieving consistent root measurements using just smartphone photos and paper. Delivering a clean, intuitive user experience without login screens or data storage. Making root trait analysis more accessible and reproducible, especially for small labs and field researchers.

What we learned

How to leverage A4 paper as a physical reference scale effectively in image analysis. Deepened our understanding of image segmentation and contour geometry. Improved performance tuning and memory handling in client-side ML apps. Learned that usability and minimalism are critical for adoption in scientific tools.

What's next for RhizoScan

Mobile camera guidance: Adding real-time feedback to help users capture better-aligned root images. Export features: Enable downloading measurements as CSV or direct upload to lab databases. Custom model training: Let users upload labeled roots to improve segmentation for different species.

Built With

• react
• tailwindcss
• tensorflow