Exploring Hijaiyyah Letter Recognition Through Teachable Machine

Abstract

Proper articulation of Hijaiyyah letters is important for Quranic recitation because precise pronunciation determines the meaning and maintains Tajweed compliance.This study investigates how Teachable Machine by Googleenables the recognition of Hijaiyyah letter pronunciation using real-time voice input. The researcher's voice provide ddata for training the model which was then evaluated using input from two different participants. The researcher recorded prediction confidence percentages for every letter and evaluated model performance through average score calculations. The letters classified into high, moderate, or low accuracy categories based on their calculated mean values.The study results show that Teachable Machine achieves reliable recognition of multiple letters especially when theletters have distinct phonetic features and clearer pronunciation. The model delivered valuable real-time feedback to learners even when certain letters presented challenges because of their similar pronunciations or softer sounds. Accessible AI tools demonstrate potential for foundational Quranic pronunciation instruction for independent learners and teacher-supplemented guidance through this exploratory study. Our findings establish a baseline for developing advanced AI-based educational toolsfor learning the Quran.