Publications by authors named "Hideo Kurozumi"
Article Synopsis
- This study evaluates the effectiveness of a fine-tuned deep convolutional neural network (CNN) in analyzing images from the pentagon copying test (PCT) to help detect dementia.
- Researchers fine-tuned and compared several pre-trained CNN models, gathering a dataset of over 1,700 PCT images from dementia-suspected patients at a hospital.
- The results showed that the GoogLeNet CNN model correctly identified PCT images with a high accuracy (area under the curve of 0.931), suggesting it could simplify the dementia screening process for healthcare professionals.
Background: There is a need for a large-scale screening test that can be used to detect dementia in older individuals at an early stage. Olfactory identification deficits have been shown to occur in the early stages of dementia, indicating their usefulness in screening tests. This study investigated the utility of an olfactory identification test as a screening test for mild cognitive dysfunction in community-dwelling older people.
View Article and Find Full Text PDF