干旱区生态环境知识资源中心(Arid): Alzheimer's Disease Prediction Using EfficientNet and Fastai

Arid

DOI	10.1007/978-3-030-82147-0_37
	Alzheimer's Disease Prediction Using EfficientNet and Fastai
	Kadri, Rahma; Tmar, Mohamed; Bouaziz, Bassem
通讯作者	Kadri, R (corresponding author), Univ Sfax Tunisia, Higher Inst Comp Sci & Multimedia, Sakiet Ezzit, Tunisia.
会议名称	14th International Conference on Knowledge Science, Engineering, and Management (KSEM)
会议日期	AUG 14-16, 2021
会议地点	Tokyo, JAPAN
英文摘要	Deep Learning has shown promising results on the field of Alzheimer's computerized diagnosis based on the neuroimaging data. Alzheimer disease is an irreversible and progressive neurodegenerative disorder that destroys gradually the brain cells. This chronic disease affect the ability of the person to carry out daily tasks. It caused many problems such as cognitive deficits, problem with recognition, memory loss and difficulties with thinking. The major breakthrough in neuroscience today is the accurate early detection of the Alzheimer's disease based on various brain biomarkers. Magnetic resonance imaging (MRI) is a noninvasive brain modality widely used for brain diseases detection specifically Alzheimer disease. It visualize a discriminate feature of the neurodegeneration which is the progressive cerebral atrophy. Various studies based on deep learning models have been proposed for the task of Alzheimer's disease classification and prediction from brain MRI scans. However these models have been implemented from scratch. The training from scratch is tedious and time-consuming task. In this paper we conduct an analysis using Open Access Series of Imaging Studies (OASIS) dataset based on tuning different convolution neural networks. Further we propose a uni-modal Alzheimer method prediction using Efficientnet network. The Efficientnet network solve the main issues of the existing convolution neural network. Data preparation includes some preprocessing steps such as image resizing. We achieve 79% using VGG16, 92% using Resenet,93% using Densenet and we produce an accuracy of 96% using the Efficientnet model.
英文关键词	Alzheimer's prediction CNN Transfer learning Fast.ai Efficientnet network
来源出版物	KNOWLEDGE SCIENCE, ENGINEERING AND MANAGEMENT, KSEM 2021, PT II
ISSN	0302-9743
EISSN	1611-3349
出版年	2021
卷号	12816
页码	452-463
ISBN	978-3-030-82147-0; 978-3-030-82146-3
出版者	SPRINGER INTERNATIONAL PUBLISHING AG
类型	Proceedings Paper
语种	英语
收录类别	CPCI-S
WOS记录号	WOS:000691057400037
WOS类目	Computer Science, Artificial Intelligence
WOS研究方向	Computer Science
资源类型	会议论文
条目标识符	http://119.78.100.177/qdio/handle/2XILL650/365615
作者单位	[Kadri, Rahma; Tmar, Mohamed; Bouaziz, Bassem] Univ Sfax Tunisia, Higher Inst Comp Sci & Multimedia, Sakiet Ezzit, Tunisia
推荐引用方式 GB/T 7714	Kadri, Rahma,Tmar, Mohamed,Bouaziz, Bassem. Alzheimer's Disease Prediction Using EfficientNet and Fastai[C]:SPRINGER INTERNATIONAL PUBLISHING AG,2021:452-463.