AUDIO AND FACIAL FEATURE ANALYSIS FOR SEPARATING DEPRESSION AND DEMENTIA
Journal: International Journal of Advanced Research in Engineering and Technology (IJARET) (Vol.11, No. 08)Publication Date: 2020-08-31
Authors : Brian Sumali; Yasue Mitsukura Taishiro Kishimoto; Masaru Mimura;
Page : 532-539
Keywords : facial landmarks; feature analysis; machine learning; pseudodementia; speech analysis.;
Abstract
Dementia is a general term for loss of cognitive ability such as memory, language, logic. It is commonly caused by neurological disorders. However, depression, a type of mental health disorder, may also cause dementia-like symptoms. When a person is afflicted with temporary cognitive impairment caused by mental health disorders, that person is said to be afflicted with pseudodementia. It differs from true dementia as the cognitive deficits are gone when the mental health disorders are cured. The most common cause, if not all, of pseudodementia cases is depression, and the risk of comorbidity of dementia with depression in elderly patients deludes even expert psychologists. Pseudodementia is typically can be diagnosed with extensive testing, but it is time consuming and taxing for both the psychiatrists and the patient. When a psychiatrist cannot definitively screen a pseudodementia case, the patient is subjected to both depression screening and dementia screening. Additionally, although machine learning has been utilized in automated mental health screening, attention for pseudodementia is minimal. In this research we aim to extract both audio features and facial features from actual dementia patients and depression patients, which were conducted and diagnosed by clinical psychiatrists. The features extracted from dementia patients and depression patients served as the basis for screening pseudodementia. We also tried to utilize machine learning using facial features and audio features to examine the possibility of automated pseudodementia screening. As a result, 85.8% accuracy was obtained by utilizing only audio features and 86.2% accuracy was obtained with only facial features. Although this can be considered satisfactory, improvement must be made before implementation in actual clinical scene is permitted.
Other Latest Articles
- BILINEAR ISOTROPIC AND BILINEAR KINEMATIC HARDENING OF AZ31 MAGNESIUM ALLOY
- MACHINE LEARNING APPROACH TO CLASSIFY THE SENTIMENT VALUE OF NATURAL LANGUAGE PROCESSING IN TELUGU DATA
- DNN BASED PET IDENTIFICATION FOR HETEROGENEOUS EDGE DEVICES
- ANALYSIS OF HYDRODYNAMIC BEHAVIOUR OF BRAHMANI RIVER USING HEC-RAS
- EFFECT ON CBR VALUE OF EXPANSIVE SOIL WITH GEOGRID REINFORCEMENT
Last modified: 2021-02-20 15:12:32