ML · Computer Vision · Tool
HatSizer
A playful yet practical tool powered by machine learning and computer vision, designed to estimate your hat size.
Overview
HatSizer uses computer vision and neural networks to measure a person's head and return an accurate hat size estimate, no tape measure needed. Built as part of NYU's Physical Computing program, it bridges the gap between hardware interaction and machine learning inference.
Problem
Buying hats online is a guessing game. Most people don't know their hat size, and sizing guides require measuring tools most don't have on hand. HatSizer makes this instant and accessible using just a camera.
Approach
The tool captures an image, runs it through a computer vision pipeline to detect and measure key facial landmarks, and maps those measurements to standard hat sizing. The neural network was trained and tuned as part of the NYU ITP winter show.
Project update in progress...