- Title
- Handwriting dynamics assessment using deep neural network for early identification of Parkinson's disease
- Creator
- Kamran, Iqra; Naz, Saeeda; Razzak, Imran; Imran, Muhammad
- Date
- 2021
- Type
- Text; Journal article
- Identifier
- http://researchonline.federation.edu.au/vital/access/HandleResolver/1959.17/185384
- Identifier
- vital:16667
- Identifier
-
https://doi.org/10.1016/j.future.2020.11.020
- Identifier
- ISBN:0167-739X (ISSN)
- Abstract
- The etiology of Parkinson's disease (PD) remains unclear. Symptoms usually appear after approximately 70% of dopamine-producing cells have stopped working normally. PD cannot be cured, but its symptoms can be managed to delay its progression. Evidence suggests that early diagnosis is important in establishing an effective pathway for management of symptoms. However, PD diagnosis is challenging, particularly in the early stages of the disease. In this paper, we present a method for early diagnosis of PD using patients’ handwriting samples. To improve performance, we combined multiple PD handwriting datasets and used deep transfer learning-based algorithms to overcome the challenge of high variability in the handwritten material. Our approach achieved excellent PD identification performance with 99.22% accuracy on illuminated task of combined HandPD, NewHandPD and Parkinson's Drawing datasets, demonstrating the superiority of our approach over current state-of-the-art methods. © 2020 Elsevier B.V.
- Publisher
- Elsevier B.V.
- Relation
- Future Generation Computer Systems Vol. 117, no. (2021), p. 234-244
- Rights
- All metadata describing materials held in, or linked to, the repository is freely available under a CC0 licence
- Rights
- Copyright © 2020 Elsevier B.V.
- Subject
- 4606 Distributed computing and systems software; 4609 Information systems; 4605 Data management and data scienceBrain disorder; Neurological disorder; Parkinson's; PD identification; Transfer learning
- Reviewed
- Funder
- The authors thank the Deanship of Scientific Research and RSSU at King Saud University for their technical support. Imran’s work is also supported by the qa through research group project number RG-1435-051.
- Hits: 500
- Visitors: 349
- Downloads: 0
Thumbnail | File | Description | Size | Format |
---|