HDFT-MViT: A Progressive Core-Enhanced Mix Framework for Alzheimer's Disease Classification using MRI images
-
By
-
Zhang, Dongyan
-
Zhang, Jincan
-
Liu, Bo
-
Liu, Min
-
Chen, Wenna
-
Du, Ganqin
-
May 31, 2026
-
Clinical Scorecard: HDFT-MViT: An Advanced Core-Enhanced Mixed Framework for Classifying Alzheimer's Disease through MRI Imaging
At a Glance
| Category | Detail |
|---|
| Condition | Alzheimer's Disease (AD) |
| Key Mechanisms | Hybrid architecture integrating hierarchical dynamic filter and lightweight Transformer for feature extraction and modeling. |
| Target Population | Patients with Alzheimer's Disease |
| Care Setting | Clinical settings utilizing MRI for diagnosis |
Key Highlights
- Proposes HDFT-MViT, a lightweight hybrid architecture based on MobileViT.
- Achieves state-of-the-art classification accuracies of 98.85% on ADNI-1 and 98.07% on ADNI-2 datasets.
- Utilizes a progressive Core-Enhanced Mix design for efficient feature extraction.
- Incorporates a channel attention mechanism to enhance feature discriminability.
- Maintains a lightweight profile with only 3.46 M parameters.
Guideline-Based Recommendations
Diagnosis
- Utilize MRI-based computer-aided diagnosis for early detection of Alzheimer's Disease.
Management
- Implement HDFT-MViT for improved classification accuracy in clinical settings.
Monitoring & Follow-up
- Evaluate model performance on public datasets like ADNI-1 and ADNI-2.
Risks
- Consider computational complexity in resource-constrained settings.
Patient & Prescribing Data
Individuals diagnosed with Alzheimer's Disease.
HDFT-MViT offers a promising tool for clinical diagnosis.
Clinical Best Practices
- Adopt hybrid models for enhanced diagnostic accuracy in Alzheimer's Disease.
- Leverage MRI imaging for comprehensive assessment.
Related Resources & Content
