３次元畳み込みニューラルネットワークを用いた単一モーダルマルチタスク学習による認知機能検査スコア推定の検討

DOI HANDLE

書誌事項

タイトル別名
  • STUDY OF ESTIMATION OF COGNITIVE ASSESMENT SCORES BY SINGLE MODAL MULTITASK LEARING USING 3 DIMENSIONAL CONVOLUTIONAL NEURAL NETWORKS

この論文をさがす

抄録

Quantitatively measuring the progress of cognitive decline to clarify Mild Cognitive Impairment (MCI) patient has received significant attention in the field of Computer Vision for Medical Imaging. Recent studies have shown promising results for an automated system to estimate various cognitive assessment scores. However, previous models are implemented in such manner that the model accurately estimates scores based on neuroanatomical visual features and measurements that are manually crafted from the original imaging modality during the preprocessing stage. To the best of our knowledge, there are no known estimators that use raw 3D voxel image as an input. For a deeper understanding of early stages of dementia, we provide the basis of an interpretable deep learning model, by implementing a basic 3D Convolutional Neural Network (CNN) model that accurately estimates the Alzheimer's Disease Assessment Scale (ADAS) scores. Based on 10-fold cross validation, our estimation model has achieved correlation of 0.51. For the next action, we would refine the model architecture and generate visual interpretations for evaluation.

収録刊行物

キーワード

詳細情報 詳細情報について

  • CRID
    1390572174784794240
  • NII論文ID
    120006897011
  • NII書誌ID
    AA12677220
  • DOI
    10.15002/00022892
  • HANDLE
    10114/00022892
  • ISSN
    21879923
  • 本文言語コード
    ja
  • データソース種別
    • JaLC
    • IRDB
    • CiNii Articles
  • 抄録ライセンスフラグ
    使用可

問題の指摘

ページトップへ