Scholarlink

This work shows our pipeline for performing Mexican Sign Language recognition. We show the main steps of our process, which perform well when detecting letter signing (spelling). In particular, we propose to exploit alternative strategies for features of landmarks from the hand so that a simple MLP network can perform well. This approach will allow us to transport our model to a more extensive vocabulary and devices without heavy computing.

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

Alternative Strategies for Feature Engineering in a Mexican Sign Language Recognition

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Erik Ivanov Domínguez Rivera,
  • Yael Josué Badillo Santos,
  • Efren Antonio Hernández Jiménez,
  • Aldo Santiago Alejandro,
  • Ivan Vladimir Meza Ruiz

摘要

This work shows our pipeline for performing Mexican Sign Language recognition. We show the main steps of our process, which perform well when detecting letter signing (spelling). In particular, we propose to exploit alternative strategies for features of landmarks from the hand so that a simple MLP network can perform well. This approach will allow us to transport our model to a more extensive vocabulary and devices without heavy computing.