International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 12 Issue: 09 | Sep 2025
p-ISSN: 2395-0072
www.irjet.net
ThinkLens: Providing Education Across Barriers Rachit Mittal1, Tarq Gaur2, Pranav Shankar3 1,2,3 Student, Sunbeam Group of Educational Institutions, Bhagwanpur, Varanasi, Uttar Pradesh, India
---------------------------------------------------------------------***--------------------------------------------------------------------passive consumption. ThinkLens is designed to fill this gap Abstract - ThinkLens is a portable, ruggedized, voice-first
by offering a single, rugged, offline-ready, voice-first device that bridges object-level perception and curriculum-aligned pedagogy, such that the ai makes the children rely on themselves with maximum family interaction.
learning companion that integrates real-time computer vision, edge-based natural language processing, and gamified pedagogy to enable contextual learning for children, elderly users, and people with special needs. The device addresses critical gaps in passive screen-based learning and digital literacy barriers through offlinecapable AI models optimized for educational content delivery. Our system combines object detection using quantized MobileNetV2 architectures, optical character recognition for text reading, and adaptive speech synthesis for multilingual support. The hardware design features an 8MP camera, Rasberry Pi, 5000mAh battery, and IP65 ruggedized encasing targeting a retail price of ₹5,000. Pilot studies with 180 participants across three demographic groups over a period of 2 years, demonstrate 23% improvement in learning retention, 67% increase in technology confidence among elderly users, and 78% ease of accomodation in day to day activities for specially abled poeple (primarily blind people). The device also operates primarily offline, ensuring privacy and accessibility in lowconnectivity environments while supporting 12 major Indian languages.
1.2 Research Objective This research aims to design and develop ThinkLens, an offline AI-powered learning companion that addresses creativity decline through voice-first interaction and realworld exploration. Specific objectives include: developing lightweight edge AI models for object recognition and natural language processing, creating accessible voicebased interfaces suitable for diverse age groups and abilities, implementing privacy-preserving on-device processing, and validating the system's effectiveness in promoting active learning and creativity. 1.3 Contributions This paper makes the following contributions: 1.
Key Words: ThinkLens, edge AI, computer vision, voicefirst interface, gamified learning, accessibility, offline NLP, education technology, absence of smart learning tools to assist diverse people
2.
1.INTRODUCTION 1.1 Motivation
3.
Access to interactive, contextualized learning remains uneven across socioeconomic strata and age groups. Children often learn passively from screens; elderly and special-needs users face barriers due to complex user interfaces and lack of localized content. Devices that combine real-world exploration with instant, voice-based feedback can transform informal learning into active, scaffolded experiences.
4. 5. 6.
Furthermore, children today are increasingly relying heavily on AI tools to complete their work. While AI can be a helpful assistant, this over-reliance risks diminishing creativity — the very spark that has driven revolutions and breakthroughs throughout human history. This trend represents a critical call to action: to rethink how we perceive learning and to build systems that encourage curiosity, exploration, and original thought rather than
© 2025, IRJET
|
Impact Factor value: 8.315
A comprehensive design for a portable edge-AI device (ThinkLens) combining computer vision, NLP, and gamification targeted at children, elderly, and special-needs users. A software architecture optimized for low-latency, privacy-preserving on-device inference, and modular content management to support localization and curriculum alignment. Dataset collection strategies and model training approaches suitable for constrained hardware and regional contexts (e.g., Indian languages, currency, cultural artifacts). An evaluation and user-study protocol assessing learning outcomes, accessibility, and usability across diverse user groups. Ethical, privacy, and deployment considerations for responsible use and scale. Make it cost effective
2. Related Work 2.1 Assistive Vision Device Prior devices (e.g., OrCam MyEye, eSight) focus on targeted assistive features such as text reading and facial recognition; these devices are often costly and optimized
|
ISO 9001:2008 Certified Journal
|
Page 205