Body Language and Speech Analysis Using Deep Learning for Enhanced Virtual Job Interviews

Satheesh Kumar A., Naveena Devi S., Preetha R., Subika K. V.

2025

Abstract

Virtual job interviews have become much more common now, but judging whether a candidate is confident, honest and a good communicator is difficult when done via a screen. In this paper, we present a deep learning-based multipurpose AI-enabled virtual interview assessment system with body language analysis and speech sentiment detection of the candidates along with facial expression recognition. It leverages pose estimation techniques (OpenPose, MediaPipe), CNNs (VGGFace, FaceNet), and speech processing models (MFCC, LSTM, BERT) to facilitate 360-degree, unbiased/speech-free, real-time assessment of candidate performance. Unlike traditional hiring methods, this model reduces subjectivity, increases hiring transparency, and generates real-time, explainable feedback for recruiters and candidates alike. The proposed solution use of privacy-preserving AI techniques, compliance to ethical standards (GDPR, CCPA) and integration with HR systems to make hiring fair, scalable and future-ready. This research establishes a new standard for AI-driven, data-informed virtual hiring by addressing certain inadequacies in existing AI-based recruitment models.

Paper Citation

in Harvard Style

A. S., S. N., R. P. and V. S. (2025). Body Language and Speech Analysis Using Deep Learning for Enhanced Virtual Job Interviews. In Proceedings of the 1st International Conference on Research and Development in Information, Communication, and Computing Technologies - ICRDICCT`25; ISBN 978-989-758-777-1, SciTePress, pages 219-225. DOI: 10.5220/0013880400004919

in Bibtex Style

@conference{icrdicct`2525,
author={Satheesh A. and Naveena S. and Preetha R. and Subika V.},
title={Body Language and Speech Analysis Using Deep Learning for Enhanced Virtual Job Interviews},
booktitle={Proceedings of the 1st International Conference on Research and Development in Information, Communication, and Computing Technologies - ICRDICCT`25},
year={2025},
pages={219-225},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0013880400004919},
isbn={978-989-758-777-1},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 1st International Conference on Research and Development in Information, Communication, and Computing Technologies - ICRDICCT`25
TI - Body Language and Speech Analysis Using Deep Learning for Enhanced Virtual Job Interviews
SN - 978-989-758-777-1
AU - A. S.
AU - S. N.
AU - R. P.
AU - V. S.
PY - 2025
SP - 219
EP - 225
DO - 10.5220/0013880400004919
PB - SciTePress