Quantifying the impact of Pyramid Squeeze Attention mechanism and filtering approaches on Alzheimer's disease classification

• Published in: Computers in biology and medicine Vol. 148; p. 105944
• Main Authors: Yan, Bin, Li, Yang, Li, Lin, Yang, Xiaocheng, Li, Tie-qiang, Yang, Guang, Jiang, Mingfeng
• Format: Journal Article
• Language: English
• Published: United States Elsevier Ltd 01.09.2022; Elsevier Limited
• Subjects: Alzheimer Disease; Alzheimer's disease; Anisotropy; Brain; Brain research; Classification; Datasets; Deep learning; Dementia; Diffusion; Fully convolutional network; Heat conductivity; Humans; Image classification; Image filtering; Image filters; Image processing; Machine learning; Magnetic Resonance Imaging; Medical imaging; Medical research; Medicin och hälsovetenskap; Mental Status and Dementia Tests; Multilayer perceptrons; Multilayers; Neighborhoods; Neural networks; Neural Networks, Computer; Neurodegenerative diseases; Neuroimaging; Noise; Pyramid squeeze attention mechanism; Teaching methods; Wavelet transforms; Image filtering; Pyramid squeeze attention mechanism; Fully convolutional network; Alzheimer's disease; Image classification
• ISSN: 0010-4825; 1879-0534; 1879-0534
• DOI: 10.1016/j.compbiomed.2022.105944

Brain medical imaging and deep learning are important foundations for diagnosing and predicting Alzheimer's disease. In this study, we explored the impact of different image filtering approaches and Pyramid Squeeze Attention (PSA) mechanism on the image classification of Alzheimer's disease. First, during the image preprocessing, we register MRI images and remove skulls, then apply median filtering, Gaussian blur filtering, and anisotropic diffusion filtering to obtain different experimental images. After that, we add the Squeeze and Excitation (SE) mechanism and Pyramid Squeeze Attention (PSA) mechanism to the Fully Convolutional Network (FCN) model respectively, to obtain each MRI image's corresponding feature information of disease probability map. Besides, we also construct Multi-Layer Perceptron (MLP) model's framework, combining feature information of disease probability map with age, gender, and Mini-Mental State Examination (MMSE) of each sample, to get the final classification performance of model. Among them, the accuracy of the MLP-C model combining anisotropic diffusion filtering with the Pyramid Squeeze Attention mechanism can reach 98.85%. The corresponding quantitative experimental results show that different image filtering approaches and attention mechanisms provide effective assistance for the diagnosis and classification of Alzheimer's disease. The overall workflow of our proposed Alzheimer's disease identification technique. (A) Image Preprocessing. After original image registration, we remove irrelevant skull parts. Then, we obtain different experimental images through median filtering, Gaussian blur filtering, and anisotropic diffusion filtering. (B) Model Training and Classification. After FCN model training, we can get MCC heatmap and feature information of disease probability map. In addition, we select the discriminative brain regions (regions of interest) for MLP model according to MCC heatmap, and combine feature information of disease probability map with age, gender, MMSE to classify experimental images after image preprocessing. [Display omitted] •Proposed a new method for image classification of Alzheimer's disease by using the Pyramid Squeeze Attention (PSA) mechanism and Fully Convolutional Network (FCN).•Quantified the impact of different image filtering approaches on classification model performance.•Three different Multi-Layer Perceptron (MLP) models were proposed to improve the classification performance of AD.