EUVIMEDEUVIMED entdecken
Crossref

Multimodal Depression Detection An Integrated Multimodal Framework for Automated Depression Severity Assessment

Anuj Sahu, Chandra Singar, Puja Gupta, Divyansh Ganote, Sanjay Patil

International Journal of Leading Research Publication · 14.05.2026 · 7 (5)

DOI: 10.70528/ijlrp.v7.i5.2191 ↗

Kurz gefasst

Fachartikel in der Zeitschrift International Journal of Leading Research Publication, online erschienen am 14.05.2026. Verfasst von Anuj Sahu und 4 weiteren Autor:innen.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Depression remains one of the most prevalent mental health disorders globally, yet early detection remains challenging due to its multifaceted and heterogeneous clinical presentation [1]. This study presents a novel multimodal depression detection system that integrates five distinct analytical modalities: facial expression analysis, vocal tone assessment, video-based emotion recognition, text sentiment analysis, and clinical questionnaire evaluation. Using a probability-weighted scoring mechanism combined with an unequal weighting scheme that prioritizes explicit user input, the system generates a comprehensive depression severity score on a 0–100 scale with clinically relevant classification thresholds. The architecture leverages convolutional neural networks (CNNs) [4] for image and video processing, long short-term memory (LSTM) networks [5] for audio analysis, a trained sequence model for text classification, and large language model (LLM) integration via the Groq API for open-ended questionnaire assessment. By combining noisy ambient snapshots with explicit clinical input, the system achieves a holistic psychological profile that extends beyond single-modality approaches [8]. This paper describes the technical implementation, mathematical framework, and clinical classification methodology underlying this integrated system.

Zitieren

APA

Sahu, A., Singar, C., Gupta, P., Ganote, D., & Patil, S. (2026). Multimodal Depression Detection An Integrated Multimodal Framework for Automated Depression Severity Assessment. International Journal of Leading Research Publication, 7(5). https://doi.org/10.70528/ijlrp.v7.i5.2191

Vancouver

Sahu A, Singar C, Gupta P, Ganote D, Patil S. Multimodal Depression Detection An Integrated Multimodal Framework for Automated Depression Severity Assessment. International Journal of Leading Research Publication. 2026;7(5). doi:10.70528/ijlrp.v7.i5.2191

BibTeX
@article{Sahu2026,
  author = {Sahu, Anuj and Singar, Chandra and Gupta, Puja and Ganote, Divyansh and Patil, Sanjay},
  title = {Multimodal Depression Detection An Integrated Multimodal Framework for Automated Depression Severity Assessment},
  journal = {International Journal of Leading Research Publication},
  year = {2026},
  volume = {7},
  number = {5},
  doi = {10.70528/ijlrp.v7.i5.2191}
}
Kostenlos starten →