Grid Slicer

Web Application
Finished Product
Back-end Code

Inspiration

As a 2D game maker, I got tired of downloading sprite sheets off the internet with messy backgrounds or fake "transparent" checkerboard patterns. Manual cropping takes way too long, so I wanted to create a tool that cuts them out automatically.

What it does

Grid Slicer takes any 2D sprite sheet image (PNG, JPG, WEBP) and cuts out each character frame into clean, separate images.

Auto Detection: Finds character shapes and extracts them into separate files.
Ordered Frames: Saves every frame in order (frame_01.png, frame_02.png).
Smart Grid Fallback: If an image has a fake checkerboard background, it neatly slices it using a clean grid.

How I built it

AI Assistance: I directed the project, logic, and design while using AI tools to write the Python and web application code.
Python & OpenCV: Created the core computer vision logic to detect shapes, handle pixel channels, and slice sprite sheets.
Web Interface: Built an interactive web application so anyone can drag and drop images directly in their browser and download frame ZIPs.

Challenges I ran into

Dealing with fake "transparent" checkerboard PNGs from search engines. The alternating gray and white squares kept tricking the detector. I fixed this by implementing a grid fallback system whenever a fake background is detected.

Accomplishments that I'm proud of

Going from zero coding experience to directing, building, and fully understanding a complete computer vision tool and web application that solves a real pain point for game creators!

What I learned

How computer vision handles image channels and bounding boxes behind the scenes, how to debug edge cases in image processing, and how to use AI as a programming collaborator.