Presentation
You should be logged in to see the presentation
Technical Session
CIBB 2021 - Neurodegenrative Diseases Session
End-to-end facial landmark detection to characterise oro-facial impairments in neurological patients: towards innovative techniques for the assessment of dysarthria
Lucia Migliorelli
Patients suffering from neurodegenerative diseases (e.g., amyotrophic lateral sclerosis (ALS) and stroke) may develop dysarthria, a neurological disorder that causes pervasive and unrelenting speech troubles. Dysarthria-evolution monitoring, via the assessments of oro-facial musculature, is crucial for allowing clinicians to promptly implement corrective and compensatory strategies to tackle communicative disability. To support clinicians, we propose a deep learning-based approach to facial landmarks detection as useful prior for assessing the oro-facial muscles impairments caused by ALS and stroke. When tested on the Toronto NeuroFace dataset, which is the largest publicly available annotated dataset in the field, the proposed approach reached a normalized mean error equal to 1.79, promptly the possibility of translating such technology in the actual clinical practice.
Lucia Migliorelli
Università Politecnica delle Marche, Italy
There are still no questions. Be the first to post one!
You should be logged in order to use this feature.
