Scholarlink

This study presents a novel approach for recognizing and translating Tamil fingerspelling using image-to-text conversion and subsequent text-to-speech synthesis. Focusing on the Tamil Language Finger Spelling 23 (TLFS 23) dataset, which contains 247 characters depicting a wide array of Tamil fingerspelling gestures, we apply a multiclass classification process to develop an effective model for accurately identifying and transcribing images of fingerspelling into their corresponding textual forms. Following transcription, the text is converted into spoken Tamil through speech synthesis techniques to improve communication accessibility for Tamil-speaking individuals. The comprehensive evaluation conducted on the TFLS 23 dataset supports the method’s viability and effectiveness, significantly mitigating language barriers and enhancing communication within the Tamil fingerspelling community.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Real-Time Tamil Fingerspelling Translation System: Image-to-Text Conversion and Speech Synthesis

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Harith Sai Saraf,
  • Sankalp Karthi,
  • Sanjay Krishna Rajmohan Menon,
  • R. T. Sadhindran,
  • G. Jyothish Lal

摘要

This study presents a novel approach for recognizing and translating Tamil fingerspelling using image-to-text conversion and subsequent text-to-speech synthesis. Focusing on the Tamil Language Finger Spelling 23 (TLFS 23) dataset, which contains 247 characters depicting a wide array of Tamil fingerspelling gestures, we apply a multiclass classification process to develop an effective model for accurately identifying and transcribing images of fingerspelling into their corresponding textual forms. Following transcription, the text is converted into spoken Tamil through speech synthesis techniques to improve communication accessibility for Tamil-speaking individuals. The comprehensive evaluation conducted on the TFLS 23 dataset supports the method’s viability and effectiveness, significantly mitigating language barriers and enhancing communication within the Tamil fingerspelling community.