Related Experiment Video
Updated: Sep 18, 2025
09:47
Author Spotlight: Advancing Alzheimer's Research – Exploring Early Detection and Multi-Omics Approaches
Published on: December 15, 2023
1.3K
RESIGN: Alzheimer's Disease Detection Using Hybrid Deep Learning based Res-Inception Seg Network
K Amsavalli1, S Kanaga Suba Raja2, S Sudha3,4
1Department of Artificial Intelligence and Data Science, Easwari Engineering College, Chennai, Tamil Nadu, India.
Current Alzheimer Research
|June 20, 2025
Summary
This study introduces the RESIGN deep learning model for early Alzheimer's disease (AD) detection using MRI scans. RESIGN achieves high accuracy, offering a reliable tool for clinical diagnosis and improved patient outcomes.
Area of Science:
- Medical Imaging
- Artificial Intelligence
- Neuroscience
Background:
- Alzheimer's disease (AD) poses a significant health challenge, necessitating early detection for improved patient survival rates.
- Traditional diagnostic methods face limitations in accurately identifying early-stage AD due to the brain's complexity.
- Deep learning (DL) offers a promising avenue for enhancing the accuracy and efficiency of AD diagnosis.
Purpose of the Study:
- To propose and evaluate a novel deep learning model, RESIGN (Res-InceptionSeg), for the early detection of Alzheimer's disease using MRI images.
- To leverage a hybrid approach combining segmentation and classification for robust AD diagnosis.
- To assess the performance of RESIGN against existing deep learning models in identifying AD, Mild Cognitive Impairment (MCI), and normal cognitive states.
Main Methods:
- MRI images were pre-processed using a Non-Local Means (NLM) filter to reduce noise.
- A ResNet-LSTM model was employed for feature extraction from White Matter (WM), Grey Matter (GM), and Cerebrospinal Fluid (CSF).
- An Inception V3-based classifier and SegNet were utilized for classification and abnormal brain region segmentation, respectively.
Main Results:
- The RESIGN model achieved a high accuracy of 99.46%, with a specificity of 98.68%, precision of 95.63%, recall of 97.10%, and an F1 score of 95.42%.
- RESIGN demonstrated superior performance compared to several established deep learning models including ResNet, AlexNet, DenseNet, and LSTM.
- The model showed significant accuracy improvements over ResNet18, CLSTM, VGG19, and CNN.
Conclusions:
- The RESIGN model's integration of spatial-temporal feature extraction, hybrid classification, and deep segmentation provides a highly reliable method for early AD detection.
- Robustness was confirmed through 5-fold cross-validation, with performance exceeding existing models on the ADNI dataset.
- RESIGN offers a dependable tool for clinical diagnosis, though potential dataset bias and generalizability limitations warrant further investigation.
