A Convolutional Neural Network Study on Depression and Eye Blink Analysis
Bryan G. Dadiz1, Alfio I. Regla2
1Bryan G. Dadiz, College of Computer Studies, Technological Institute of the Philippines, Manila, Philippines.
2Alfio I. Regla, College of Computer Studies, Technological Institute of the Philippines, Manila, Philippines.
Manuscript received on 04 March 2023 | Revised Manuscript received on 16 March 2023 | Manuscript Accepted on 15 April 2023 | Manuscript published on 30 April 2023 | PP: 7-11 | Volume-12 Issue-5, April 2023 | Retrieval Number: 100.1/ijitee.E94880412523 | DOI: 10.35940/ijitee.E9488.0412523
Open Access | Editorial and Publishing Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: This study aims to determine the correlation between human blinks and depression. The study utilises a convolutional neural network to detect blinks in a video. Using the Closed Eyes in the Wild dataset, the Convolutional Neural Network model was trained, achieving 99.24% training accuracy and a loss of 0.0275 after 50 epochs. However, the results from the model validation showed that 61.09% of the data from the two datasets, which were labelled using the BDI-II depression scale, were tested. The study collated the results of recorded blinks from the video datasets and found a weak positive correlation between the recorded blinks, computed as blinking rates, and depression. The result showed that the R2 score was 0.34; thus, there is a possibility, but not a highly indicative indicator of depression.
Keywords: AVEC’ 14 Introduction, Convolution Neural Network, Depression Analysis, Eye Blinks, Major Depressive Disorder, Non-Verbal Behavior, Visual Cue
Scope of the Article: Visual Analytics