Abstract:
The Zhuang people distribute in Guangxi province maintaining their language without characters. To convert the speech of the Zhuang into Chinese characters and translate ...Show MoreMetadata
Abstract:
The Zhuang people distribute in Guangxi province maintaining their language without characters. To convert the speech of the Zhuang into Chinese characters and translate it into other languages, we built a database to construct an end-to-end automatic speech recognition using a neural network SaformerNet. The network used a transformer-based deep learning model for the translation. An edge computing-based private cloud system using FPGA acceleration was proposed for the rapid response of the network. The results demonstrated that the proposed solution was suitable for small-scale conversion of natural languages in a real-time application.
Date of Conference: 11-13 August 2023
Date Added to IEEE Xplore: 04 December 2023
ISBN Information: