• Md Jishan
• Sadab Ansari
• Gopal Kumar
• Alfiya Parveen
• Binod Kumar

Mental health disorders require reliable, scalable, and objective detection. Traditional diagnosis is often subjective. We propose a new approach: Leveraging Multi-Model Deep Learning to improve mental health estimation. Our framework integrates three behavioral modalities: text, voice tone, and facial expressions. This provides a comprehensive psychological assessment. The system uses specific deep neural networks: Transformers for text, RNNs for voice, and CNNs for faces. The core innovation is a fusion mechanism that combines these features. This multimodal fusion captures subtle, crucial cues, outperforming single-modality methods. Our model classifies conditions across seven categories: Normal, Depression, Anxiety, Suicidal ideation, Bipolar disorder, Stress, and Personality disorder. Results show our multi-model architecture significantly increases accuracy and robustness. This research offers a strong basis for automated, reliable mental health monitoring tools.

Multi-model learning, mental health estimation, text analysis, voice tone, facial expression recognition, deep learning

"Leveraging Multi-Model Learning for Reliable Mental Health Estimation ", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.12, Issue 11, page no.f656-f663, November-2025, Available :http://www.jetir.org/papers/JETIR2511598.pdf

"Leveraging Multi-Model Learning for Reliable Mental Health Estimation ", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.12, Issue 11, page no. ppf656-f663, November-2025, Available at : http://www.jetir.org/papers/JETIR2511598.pdf