Automatic evaluation of eye gestural reactions to sound in video sequences. (October 2019)

• Record Type: Journal Article
• Authors: Fernández, Alba; Ortega, Marcos; de Moura, Joaquim; Novo, Jorge; Penedo, Manuel G.
• Is Part Of: Engineering applications of artificial intelligence. Volume 85(2019)
• Page Start: 164
• Page End: 174
• Publication Date: 2019-10
• Subjects: Sound responses -- Gestural reactions -- Hearing screening -- Movement detection -- Screening methods -- Video sequences; Engineering -- Data processing -- Periodicals; Artificial intelligence -- Periodicals; Expert systems (Computer science) -- Periodicals; Ingénierie -- Informatique -- Périodiques; Intelligence artificielle -- Périodiques; Systèmes experts (Informatique) -- Périodiques; Artificial intelligence; Engineering -- Data processing; Expert systems (Computer science); Periodicals; 620.00285
• DOI: 10.1016/j.engappai.2019.06.009
• Languages: English
• ISSNs: 0952-1976

Abstract

Abstract: Hearing loss is a common disorder that often intensifies with age. In some cases, especially in the elderly population, the hearing loss may decrease in physical, mental and social well-being capacities. In particular, patients with signs of cognitive impairment typically present specific clinical–pathological conditions, which complicates the analysis and diagnosis of the type and severity of hearing loss by clinical specialists. In these patients, unconscious changes in gaze direction may indicate a certain perception of sound through their auditory system. In this context, this work presents a new system that supports clinical experts in the identification and classification of eye gestures that are associated with reactions to auditory stimuli by patients with different levels of cognitive impairment. The proposed system was validated using the public Video Audiometry Sequence Test (VAST) dataset, providing a global accuracy of 97.12% for the classification of eye gestures and a 100% for gestural reactions to auditory stimuli. The proposed system offers a complete analysis of audiometric video sequences, being applicable in daily clinical practice, improving the well-being and quality of life of the patients. Graphical abstract: