Doodle Recognition for Early English Language Learning in Preschool Children

Abstract

This work presents the design and evaluation of a Doodle Recognition System aimed at improving English vocabulary retention in preschool children aged 3 to 6. The implemented interactive web application incorporates AI-based doodle recognition using the DoodleNet model to provide realtime feedback and support fine motor skill development. Implemented with Next.js, Tailwind CSS, and Supabase, the system includes a responsive drawing interface, pre- and postassessment features, and progress tracking (https://fyp-doodlenet-with-supabase.vercel.app/). User testing with 28 children revealed measurable vocabulary improvements, particularly among 4- to 5-year-olds, with an overall post-test score increment of 11.3 %. The findings indicated that integrating visual-motor activities with vocabulary learning enhances engagement and retention, suggesting the potential of doodle recognition systems to supplement early childhood language instruction.