The Abserny team
Faculty of Specific Education , CS Department , 2025–2026
“We built it for the person who picks it up alone, in the dark, for the first time.”

Seven students. One constraint: no screen, ever.

أبصرني , See Me.

The Story

Built from a
single question

Abserny began as a graduation project in late 2024 at the Faculty of Engineering. We chose accessibility because over 2.2 billion people worldwide live with visual impairment, and AI had reached a point where it could make a real, practical difference in their daily lives.

The question that shaped every decision: what if a blind user picked up this app for the very first time, alone? Most assistive tools fail at that exact moment, they require sighted help to choose a language, read instructions, or navigate setup. We set out to eliminate that dependency completely.

The result is Abserny v3.1, a fully voice-first Android application in which no visual interaction is required at any point, from first launch through every day of use.

What We Offer

Real-time Scene Description

Powered by Gemini 2.0 Flash Lite, Abserny describes your environment in natural language, hazards first, spatially aware, in Arabic or English, at 1,430 ms median latency.

Works Offline, Always

When there's no internet, ML Kit steps in automatically. The app never fails silently, it routes to on-device detection and keeps speaking. No data connection required.

Gesture-Only Control

Five gestures replace the entire UI. Double tap to scan. Long press to repeat. Triple tap for settings. Swipe to change mode. No screen interaction, no voice commands, no menus.

Our Values

01
Accessibility First Every design decision is evaluated from the perspective of a user who cannot see the screen, including before the app has started and before a language has been chosen. Accessibility is the primary constraint, not a checklist item.
02
Voice Before Visual Audio is the primary interaction channel. The user controls everything through five gestures. No screen, no menus, no visual navigation, ever. If something requires looking, it doesn't belong in Abserny.
03
Reliable by Design A finite state machine governs every interaction, making race conditions structurally impossible. Offline detection ensures the app works without internet. Nothing depends on a premium device or a data connection.

Gesture Vocabulary

Five gestures cover every function. The vocabulary is intentionally small, a larger set would be harder to remember and more error-prone.

01
Double Tap , Scan Triggers a scan and returns a spoken description. Full-screen target with a 320 ms recognition window prevents accidental activation from a single exploratory tap.
02
Long Press , Repeat Repeats the last spoken result immediately. Isolated from taps by a 700 ms timer threshold, no confusion between the two actions.
03
Triple Tap , Settings Opens the spoken Settings overlay, navigate by swipe, confirm by double tap, close by long press. Language change and tutorial replay, entirely without visual interaction.
04
Swipe Right / Left , Change Mode Cycles through Scene, Object, Read, and People detection modes. A 60 px horizontal threshold with a 1.2:1 directional ratio prevents accidental switching during normal hand movement.

Meet the Team

Seven students from the Faculty of Engineering, Computer Science Department, Academic Year 2025–2026.

Saeed Hany

Saeed Hany

Lead Developer & AI Engineer

Architecture, gesture system, AI integration, and hooks design

Omaretooo

Omaretooo

Speech & Therapy

Without him the team could be lost in the darkness

Mohamed Eid

Mohamed Eid

Speech & Language Systems

TTS queue, bilingual support, Arabic prompt engineering

Bassant Wael

Bassant Wael

Research & Testing

User testing, gesture accuracy evaluation, accessibility validation

Rashida Hassan

Rashida Hassan

Backend & Offline Systems

ML Kit integration, AbserneyVision training pipeline, offline routing

Hana Hamed

Hana Hamed

Onboarding & UX

Voice-first onboarding design, gesture tutorial, settings overlay

Mariam Suroor

Mariam Suroor

Documentation & QA

Graduation book, latency benchmarks, quality evaluation

Join Our Journey

Abserny is an open graduation project. We welcome feedback from accessibility researchers, developers, and anyone with lived experience of visual impairment.

Contribute on GitHub