HIDE - HypomImia DetEction using lightweight architectures
Description: Several studies have demonstrated the development of neurodegenerative diseases, such as Alzheimer, Parkinson's Disease, and Amyotrophic Lateral Sclerosis, may appear around 10-20 years before their clinical symptoms. The decrease in cognitive and physical functions is considered a primary side effect, impacting the quality of life. The scientific community has employed different data sources and artificial intelligence to early detect those diseases, such as magnetic resonance imaging and computerized tomography, as well as textual information from medical reports. One crucial feature of most neurodegenerative diseases concerns facial paralysis, aka hypomimia, where some muscles are compromised as the condition goes on. However, in its early stages, the disorder has a more negligible severity, making paralysis detection not straightforward. A few recent works have studied video content to aid in detecting neurodegenerative diseases, in which the individuals are requested to perform some tasks, e.g., say some sentences and do lip movements. This proposal aims to continue research with Prof. Dinesh (Royal Melbourne Institute of Technology, Australia). We intend to advance in detecting Amyotrophic Lateral Sclerosis, aka ELA, by identifying facial paralysis in multimedia content. Another research line concerns the automatic identification of Parkinson's Disease, for individuals may also figure hypomimia. (AU)
Funding type: grant
Funder scheme: Auxílio à Pesquisa - Regular
Start date: 2024-02-01
End date: 2026-01-31
Investigator: João Paulo Papa
Institution: Universidade Estadual Paulista (UNESP). Campus de Bauru. Faculdade de Ciências (FC)
Subfield: Neurology
Field: Medicine
Domain: Health Sciences
Funded outputs count: 3
Funder award ID: 23/14197-2
OpenAlex ID: G2522220048