Numerical detection of diabetic retinopathy stages by multifractal analysis for OCTA macular images using multistage artificial neural network

El-Damrawi, G.; Zahran, M. A.; Amin, E. A.; Abdelsalam, Mohamed M.

doi:10.1007/s12652-021-03565-3

Cited by 1 publication

(1 citation statement)

References 45 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Therefore, the utilization of fractal geometry in neurosciences has been the outcome of a new trend of research focused on the analysis of the complexity of biological systems [35][36][37][38][39][40][41] . The application of the fractal dimension (FD) in investigating the clinical-pathological spectrum of neurodegenerative diseases including AD 42 .…”

Section: Very Severe Cogniɵve Decline (Very Severe Demenɵa)mentioning

confidence: 99%

A new strategy for the early detection of alzheimer disease stages using multifractal geometry analysis based on K-Nearest Neighbor algorithm

Elgammal

Zahran

Abdelsalam

2022

Sci Rep

Self Cite

View full text Add to dashboard Cite

Alzheimer's Disease (AD) is considered one of the most diseases that much prevalent among elderly people all over the world. AD is an incurable neurodegenerative disease affecting cognitive functions and were characterized by progressive and collective functions deteriorating. Remarkably, early detection of AD is essential for the development of new and invented treatment strategies. As Dementia causes irreversible damage to the brain neurons and leads to changes in its structure that can be described adequately within the framework of multifractals. Hence, the present work focus on developing a promising and efficient computing technique to pre-process and classify the AD disease especially in the early stages using multifractal geometry to extract the most changeable features due to AD. Then, A machine learning classification algorithm (K-Nearest Neighbor) has been implemented in order to classify and detect the main four early stages of AD. Two datasets have been used to ensure the validation of the proposed methodology. The proposed technique has achieved 99.4% accuracy and 100% sensitivity. The comparative results show that the proposed classification technique outperforms is recent techniques in terms of performance measures.

show abstract