TensorLip: Unveiling Conversations with Deep Learning by Harnessing TensorFlow for Lip Reading Intelligence

Author:

Nandini C 1,Sasi Kumar B 1

Affiliation:

1. Raja Rajeswari College of Engineering, Bengaluru, Karnataka, India

Abstract

“TensorLip presents a pioneering approach towards the realm of speech-recognition and communication accessibility through the fusion of deep-learning and TensorFlow technology. Our paper focuses on the advancement of a lip-reading system capable of deciphering spoken language solely from visual cues of lip movements. Leveraging the power of algorithms in deep learning, particularly tailored and optimized within the TensorFlow framework, TensorLip aims to bridge the communication gap in situations where individuals experience hearing challenges or amidst noisy surroundings where traditional audio-based methods fall short. By harnessing the vast potential of neural networks, our innovative solution promises to revolutionize the manner in which we perceive and understand spoken language, thereby enhancing inclusivity and facilitating seamless communication across diverse linguistic and auditory landscapes.”.

Publisher

Naksh Solutions

Reference14 articles.

1. [1] Enhancing Lip Reading Efficiency with Efficient-GhostNet by Gaoyan Zhang and Yuanyao Lu.

2. [2] Advancing Lip Reading: WLAS Network for Open-World Recognition by Joon Son Chung and Andrew Senior, and Oriol Vinyals, Andrew Zisserman.

3. [3] Sharma, Annu, Praveena Chaturvedi, and Shwetank Arya. "Human recognition methods based on biometric technologies." International Journal of Computer Applications 120.17 (2015).

4. [4] Evolution of Automatic Lip-Reading: A Deep-Learning Perspective by Adriana Fernandez-Lopez, Federico M. Sukno.

5. [5] Advancements in Lipreading: From Traditional to Deep-Learning Methods by Mingfeng Hao, MutallipMamut, NurbiyaYadikar, Alimjan Aysa.

同舟云学术

1.学者识别学者识别

2.学术分析学术分析

3.人才评估人才评估

"同舟云学术"是以全球学者为主线,采集、加工和组织学术论文而形成的新型学术文献查询和分析系统,可以对全球学者进行文献检索和人才价值评估。用户可以通过关注某些学科领域的顶尖人物而持续追踪该领域的学科进展和研究前沿。经过近期的数据扩容,当前同舟云学术共收录了国内外主流学术期刊6万余种,收集的期刊论文及会议论文总量共计约1.5亿篇,并以每天添加12000余篇中外论文的速度递增。我们也可以为用户提供个性化、定制化的学者数据。欢迎来电咨询!咨询电话:010-8811{复制后删除}0370

www.globalauthorid.com

TOP

Copyright © 2019-2024 北京同舟云网络信息技术有限公司
京公网安备11010802033243号  京ICP备18003416号-3