How devrajdc/mlh-peaceout-hacks works
Project Type and Purpose
• This is a purely front-end, static web application (“Peace in Space”) that uses client-side machine learning (Face API via face-api.js) to detect a user’s facial emotion in real time, and then redirects the user to a mood-specific static page with tailored content (audio, PDFs, JSON data).
• There is no backend service or database; all assets (HTML, CSS, JS, models, media) are served as static files (e.g., via Vercel or any static host).
File-Structure Analysis
• Top-level files: index.html (entry point), api.js (initialization/orchestration of face detection), package.json.
• /css, /js: UI styling and client-side libraries (jQuery, Lodash).
• /models: pre-trained weight manifests and shards for face-api.js models.
• /pages: mood pages (angry.html, happy.html, neutral.html, sad.html) and a team page.
• /src: supplementary static assets (audio clips, PDFs, JSON with motivational data).
README Insights
• Tech stack: HTML, CSS, JavaScript, face-api.js (Deep Learning in the browser), Figma (design).
• Flow described: face recognition → expression detection → redirect to mood page.