Presensi Wajah: Sistem Presensi Kuliah Berbasis Pengenalan Wajah di Peramban
Year: 2026
Type: software
Abstract: Presensi Wajah is a face-recognition attendance system for university classes, developed at Universitas IPWIJA, Jakarta, Indonesia. It consists of two repositories: app-presensi (static frontend, no build step) and api-presensi (backend on Cloudflare Workers with a D1 database).Privacy-by-designFace feature extraction (TinyFaceDetector → 68 landmarks → 128-D descriptor, via face-api.js) runs entirely in the browser.Only the 128-number descriptor is sent; images and video are never uploaded or stored. Descriptors are stored encrypted on the server.Explicit biometric consent and password confirmation are required for enrolment; users can delete their face data at any time.FeaturesStudents: enrol / re-enrol / delete face data, attend with an active liveness challenge (blink, turn left, turn right), view attendance history.Lecturers: create classes, manage members (see who has not enrolled), open/close meetings with duration and late threshold, per-meeting report, manual override (excused/sick/absent with reason), CSV export protected against formula injection.Admins: statistics, account creation, role changes, password reset, face-template reset.Technical stackFrontend: Vanilla JavaScript SPA, strict Content Security Policy, server data rendered via textContent, face-api.js loaded with Subresource Integrity.Backend: Cloudflare Workers, D1 (SQLite), session tokens, captcha on login/registration.LimitationsLiveness is evaluated on the client and therefore is not a hard security boundary. Automated tests so far use mocked face models; liveness thresholds (blink EAR, head-turn ratio) still need calibration on real devices and faces. See SECURITY.md in api-presensi.Repositorieshttps://github.com/sismadi/app-presensi https://github.com/sismadi/api-presensi
Source: Figshare
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
SDG: Quality education
Open Access status: green
OpenAlex ID: W7220951069
