2024/2025 · Onboarding · Finance Portals · Web Design
View case →
SIGN LANGUAGE TRAINER
Duration
1 Month
Project Type
Machine Learning
Accessibility Design
Web Design
Web Development
My Role
Product Champion
UX Lead
Front End Dev
Target Users
Hearing Imparied
Schools & Academys
Contribution
Product Conception
User Reasearch
Flow Design
High Fidelity Design
Front End Dev
Outcome
Identified a valid market gap
Designed and developed a working prototype
2026 targeted launch
Overview
While researching accessibility for our online school, I identified a gap in support for British Sign Language (BSL) GCSE students in the UK. Exam delays and cancellations due to limited support highlighted a pressing need. I analysed hearing-impaired population statistics and projected growth, revealing a large, underserved learner community.
To address this, I explored machine-learning approaches for recognising hand gestures from camera input and prototyped a webcam-based tool that helps learners practise BSL with real-time feedback on signing accuracy. The concept targets younger learners with a simple, engaging interface and a path toward an MVP focused on BSL fundamentals.
My Role
- End-to-end owner: Sole designer and developer across research, UX, and prototype build.
- Research: Mapped market/accessibility needs and defined problem scope and users.
- ML exploration: Evaluated hand-gesture recognition approaches in Python with computer-vision tooling.
- Prototype: Built a working camera demo with real-time detection and “hold to validate” feedback.
- UI design: Created an interface for younger learners, optimised for clarity and engagement.
- Gamification: Added points/levels to motivate progression and track learning.
- Roadmap: Planned MVP stages and iteration cycles based on user testing and feedback.
Flow Analysis
- Onboarding: Simple guidance on camera setup, framing, and how to practise.
- Gesture practice: Users mimic displayed BSL signs via webcam.
- Real-time feedback: System detects the sign and requires ~3 seconds held steady to confirm accuracy.
- Scoring & progression: Successful attempts earn points/levels and unlock new signs.
- Continuous learning: Repeat, receive feedback, and progress through structured levels to reinforce retention.
The Problem
- BSL GCSEs face delays/cancellations due to limited qualified support and resources.
- Lack of accessible, tech-driven practice tools outside the classroom.
- Large and growing hearing-impaired community with scarce interactive BSL tools.
- Existing resources lack real-time feedback, making self-correction difficult.
- Younger learners need a friendly, engaging experience to sustain motivation.
The Solution
- Applied ML-based hand-gesture recognition (Python + CV) to detect BSL signs via webcam.
- Introduced “hold to validate” to improve recognition confidence and reduce false positives.
- Designed a simple, approachable UI tailored to younger learners.
- Built gamification (scoring, levels) to encourage consistent practice.
- Focused the MVP on core BSL fundamentals with a roadmap for expansion.
- Planned iterative releases informed by user testing and feedback.
‹
›
Impact (Projected)
- Validated demand for interactive BSL learning with real-time feedback.
- Delivered a functional prototype demonstrating feasibility for BSL education.
- Secured initial stakeholder and user interest with positive early feedback.
- Established a path toward a 2026 MVP targeting young learners to improve accessibility and proficiency.
- Laid groundwork for advanced models and a growing sign vocabulary in future releases.
