Stylable is a smart outfit generator which uses machine learning, object detection, webscraping, and a custom outfit algorithm to give users ideas for their outfits.
Core Libraries - YOLO, Selenium, CustomTKinter, PIL, OpenCV
Stylable begins loading the Selenium webdriver in the background. This is to expedite the webscraping process.
The user uploads a pile of clothes, selects whether they want to see masculine or feminine outfits, and then submits. Stylable then detects which outfits are present in the image, classifying them and then inputting the present outfits into our custom algorithm. The algorithm submits all possible outfits into the Selenium webdriver, to which Stylable downloads the best fit images. Those downloaded images are then shown to the user.
This is a normalized confusion matrix for all of the classifications Stylable detects.
This is a training batch after 100 epochs of training
This is a labeled batch after 100 epochs of training
I've been a developer on ROBLOX for 8 years, working many roles on dozens of games. My favorite roles have been 3D modeling and scripting items. The most popular game I've worked on is Baseplate with a total of 483.8k+ plays. My most polished game is Outlands Border, totalling 5k+ lines of code.
I would define myself as a determinist in morality but not causality. My favorite philosophers are Kierkegaard, Descartes, Plato, and Locke.
