基于YOLOV5的手语翻译系统的研究与设计

The Research and Design of a Sign Language Translation System Based on YOLOV5

本研究旨在提高听障人士的社会融入能力,使他们能够更好地参与社交活动。为此,本文基于YOLOV5s模型设计并实现了手语翻译系统,并研究采用了数据增强技术MixUp和HSV随机增强来改进系统的训练过程,进一步提高了模型的泛化能力与鲁棒性。经过数据增强训练出的手语双向翻译系统经过测试,很好的实现了语音或文字与手语之间的双向翻译,同时操作简便且实用性强,具有良好的应用前景。

The objective of this study is to enhance the social integration capabilities of individuals with hearing impairments,enabling them to actively participate in social activities.Therefore,this paper proposes and implements a sign language translation system based on the YOLOV5s model.Data augmentation techniques such as MixUp and HSV random enhancement are utilized to optimize the training process of the system,thereby improving its generalization ability and robustness.The sign language bidirectional translation system trained using data augmentation has been successfully tested and effectively facilitates communication between speech or text and sign language.Additionally,it offers simplicity,practicality in operation while showing promising prospects for application.