EUVIMEDEUVIMED entdecken
Crossref

Deceptive Acoustics in Medicolegal and Psychological Practice: An AI-Powered Framework for Detecting Voice Manipulation in Forensic Evidence

Shikha Upadhyay, Murugan R. b

International Journal of Drug Delivery Technology · 17.06.2026 · 16 (55s)

DOI: 10.25258/ijddt.16.55s.24 ↗

Kurz gefasst

Fachartikel in der Zeitschrift International Journal of Drug Delivery Technology, online erschienen am 17.06.2026. Verfasst von Shikha Upadhyay und 1 weiteren Autor:innen.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Background

Voice manipulation technologies capable of altering perceived gender or emotional register pose acute challenges for forensic medicine, psychological testimony, and medicolegal proceedings. When a recorded voice may serve as evidence in criminal, civil, or psychiatric evaluations, the inability to detect acoustic transformation constitutes a significant gap in forensic practice.

Objectives

To develop and validate an AI-based acoustic framework for detecting two forensically distinct voice manipulation types — gender conversion and emotion conversion — with direct application to medicolegal voice authentication and psychological credibility assessment.

Methods

Using CREMA-D (7,442 utterances; 91 speakers) and RAVDESS (1,440 utterances; 24 speakers), a corpus of 13,560 samples was constructed across three conditions: original, gender-converted (WORLD vocoder), and emotion-converted (StarGAN-VC). A 74-dimensional acoustic feature vector was statistically validated; four classifiers were evaluated under speaker-stratified ten-fold cross-validation.

Results

The MLP achieved F1-scores of 0.913 ± 0.019 for gender manipulation detection (AUC = 0.971) and 0.864 ± 0.024 for emotion manipulation detection (AUC = 0.936), with balanced error rates of 7.35% and 13.45% respectively.

Conclusions

AI-based acoustic analysis reliably detects voice manipulation with direct forensic and medicolegal relevance. The feature-level findings are actionable for practitioners without specialist machine learning infrastructure.

Zitieren

APA

Upadhyay, S., & b, M. R. (2026). Deceptive Acoustics in Medicolegal and Psychological Practice: An AI-Powered Framework for Detecting Voice Manipulation in Forensic Evidence. International Journal of Drug Delivery Technology, 16(55s). https://doi.org/10.25258/ijddt.16.55s.24

Vancouver

Upadhyay S, b MR. Deceptive Acoustics in Medicolegal and Psychological Practice: An AI-Powered Framework for Detecting Voice Manipulation in Forensic Evidence. International Journal of Drug Delivery Technology. 2026;16(55s). doi:10.25258/ijddt.16.55s.24

BibTeX
@article{Upadhyay2026,
  author = {Upadhyay, Shikha and b, Murugan R.},
  title = {Deceptive Acoustics in Medicolegal and Psychological Practice: An AI-Powered Framework for Detecting Voice Manipulation in Forensic Evidence},
  journal = {International Journal of Drug Delivery Technology},
  year = {2026},
  volume = {16},
  number = {55s},
  doi = {10.25258/ijddt.16.55s.24}
}
Kostenlos starten →