Application of Deep Learning Models for Color Recognition in EEG of the Elderly (81284)

Session Information:

Friday, 24 May 2024 15:30
Session: Poster Session 1
Room: Orion Hall (5F)
Presentation Type: Poster Presentation

All presentation times are UTC + 9 (Asia/Tokyo)

Color perception plays a crucial role in stimulating the human brain. Currently, research on color recognition of brainwave signals using machine learning mostly focuses on the general population. However, this study concentrates on 17 elderly individuals, including 11 with mild cognitive impairment and 6 with sub-healthy conditions, with an average age of 80 years. Building upon existing literature on machine learning applied to color recognition in brainwaves, this study first redesigned two videos based on different RGB colors. One video used gray transitions, while the other incorporated scenic transitions between each color, aiming to explore participants' overall perception and compare their recognition accuracy after viewing. In this experiment, brainwave signals were collected using a single-channel EEG device while participants watched the videos. After preprocessing, the data were inputted into DNN and CNN models for training. The DNN model achieved a recognition accuracy of 97% for individual brainwave signals, while the CNN model achieved a higher accuracy of 99%. The overall recognition rates for the mild cognitive impairment and sub-healthy groups were 93% and 97%, respectively. When data were limited, the model's recognition results deteriorated, but introducing event-related spectral perturbation (ERSP) data processing improved the overall recognition accuracy by reducing brainwave noise. Analysis results indicated that videos with scenic elements reduced psychological stress in elderly individuals, promoting relaxation and obtaining more accurate brainwave data. In the future, applying machine learning techniques is expected to deepen the understanding of how individuals with cognitive impairments perceive and respond to different colors.

Authors:
Pei-Fen Wu, National Changhua University of Education, Taiwan
Shao-Wei Yen, National Changhua University of Education, Taiwan
Ming-Chi Shih, Dementia Care Association, Taiwan R.O.C


About the Presenter(s)
Professor PEIFEN WU is a University Associate Professor/Senior Lecturer at National Changhua University of Education in Taiwan

See this presentation on the full scheduleFriday Schedule



Conference Comments & Feedback

Place a comment using your LinkedIn profile

Comments

Share on activity feed

Powered by WP LinkPress

Share this Presentation

Posted by Clive Staples Lewis

Last updated: 2023-02-23 23:45:00