Related Experiment Video
Updated: Jul 4, 2026
08:05
Design and Analysis for Fall Detection System Simplification
Published on: April 6, 2020
LSHMMformer: An Intelligent detection model of depression based on multi-modal fusion
Yuntao Shi1, Tian Gan1, Jie Li1
1North China University of Technology, 5 Jinyuanzhuang Road, Beijing, 100144, PR China.
Journal of Neuroscience Methods
|April 9, 2026
Summary
A new deep learning model, LSHMMformer, enhances depression detection accuracy and efficiency. This intelligent system aids in personalized screening and early intervention for mental well-being.
Area of Science:
- Artificial Intelligence
- Computational Psychiatry
- Machine Learning
Background:
- Depression poses a significant threat to mental well-being globally.
- Current automatic depression detection methods struggle with efficiency and accuracy.
- Deep learning (DL) offers potential for cost-effective diagnosis and personalized treatment.
Purpose of the Study:
- To develop an intelligent depression detection model.
- To improve prediction performance and enhance detection efficiency.
- To address limitations in current depression detection methodologies.
Main Methods:
- Proposed an intelligent depression detection model named LSHMMformer.
- Integrated a novel Locality-Sensitive Hashing Attention (LSHAttention) mechanism with a learnable projection matrix.
- Employed multi-modal fusion using stacked Transformer Decoder layers for audio and text data.
Main Results:
- Tested on the EATD-Corpus dataset.
- Achieved a mean absolute error of 5.66.
- Achieved a root mean square error of 8.68.
Conclusions:
- The LSHMMformer model significantly improves depression detection accuracy.
- Facilitates personalized depression screening software development.
- Reduces clinical diagnostic barriers and supports early intervention decisions.
