Mahalakshmi S, Valarmathi K

AD-HOLDER: Alzheimer's Disease Detection via Machine Learning based Histograms of Light GBM Classifier using MRI and PET Images

Introduction

Ad-holder: alzheimer's disease detection via machine learning based histograms of light gbm classifier using mri and pet images. AD-HOLDER: Early Alzheimer's disease detection using a novel machine learning model with MRI & PET images. Achieves 99.12% accuracy, outperforming existing GBA, XAI, CAD.

Abstract

Alzheimer's disease (AD) is an irreversible brain condition that impairs memory and cognitive processes. However, existing Alzheimer detection methods have shown low diagnostic accuracy (ACC) due to limited images and inefficient feature analysis. In this paper, a novel AD-HOLDER model is proposed for early recognition of AD using dual imaging methods: magnetic resonance imaging (MRI) and positron emission tomography (PET). The proposed AD-HOLDER model presents an integrated framework that uniquely combines Deep Image Prior (DIP) denoising, Histogram of Oriented Gradients (HOG) feature extraction, and Light Gradient Boosting Machine (LGBM) classification for AD detection from MRI and PET images. The HOG method aims to enhance the spatial and contextual representation of neurological patterns by combining structural features from MRI and statistical features from PET images. A classifier based on the LGBM processes the dual features to classify images as either normal or abnormal, effectively capturing complex patterns and improving classification ACC. The abnormal region is segmented using a graph-based segmentation (GBS) model to accurately detect affected areas for accurate detection of AD. The effectiveness of the proposed AD-HOLDER model is evaluated using ACC, specificity (SPE), precision (PRE), recall (REC), and F1-score (F1) based on the OASIS dataset. The proposed AD-HOLDER model achieves a classification ACC of 99.12 % through machine learning. The proposed AD-HOLDER model increases overall ACC by 1.55 %, 25.44 %, and 3.14 % compared to the Gradient Boosting Algorithm (GBA), Explainable Artificial Intelligence (XAI), and Computer-Aided Diagnosis (CAD) systems, respectively.

Review

The paper "AD-HOLDER: Alzheimer's Disease Detection via Machine Learning based Histograms of Light GBM Classifier using MRI and PET Images" introduces a novel AD-HOLDER model aiming to improve the diagnostic accuracy of Alzheimer's disease (AD) detection. The authors identify critical limitations in current methods, particularly regarding low accuracy stemming from limited image data and inefficient feature analysis. To address this, their proposed AD-HOLDER model presents an integrated framework that uniquely combines Deep Image Prior (DIP) denoising, Histogram of Oriented Gradients (HOG) for feature extraction, and a Light Gradient Boosting Machine (LGBM) for classification, specifically leveraging dual imaging modalities: MRI and PET. A significant strength of the AD-HOLDER model lies in its multi-component and multi-modal approach. The combination of DIP denoising to improve image quality, followed by HOG feature extraction to consolidate structural information from MRI and statistical features from PET, provides a rich representation for the classifier. The use of LGBM for classification, known for its efficiency and accuracy in handling complex patterns, is a sound choice for distinguishing between normal and abnormal images. Furthermore, the integration of a graph-based segmentation (GBS) model to precisely locate affected regions enhances the interpretability and clinical utility of the detection. The reported high classification accuracy of 99.12% on the OASIS dataset, alongside significant performance gains over comparison methods like GBA, XAI, and CAD systems, underscores the potential effectiveness of this integrated methodology. While the reported accuracy is remarkably high and promising, it also prompts a need for further investigation into the specific conditions and potential generalizability of these results. The abstract does not explicitly detail the classification task (e.g., healthy vs. AD, or including mild cognitive impairment stages), which is crucial for assessing its clinical relevance and potential impact on early diagnosis. An accuracy of 99.12%, while impressive, often warrants scrutiny regarding potential overfitting or the specific characteristics and preprocessing of the dataset. Future work would benefit from validation on more diverse, larger, and multi-site datasets to confirm robustness and generalizability. Additionally, clarifying the title's "Histograms of Light GBM Classifier" phrasing, perhaps by more explicitly stating "HOG features with Light GBM Classifier," could enhance clarity for the reader.

Full Text

Comments

You need to be logged in to post a comment.