FSCSCOOT: Functional Calculus Competitive Swarm Coot Optimization-based CNN transfer learning for Parkinson’s disease classification

Abstract

Parkinson's disease (PD) is a neurological disorder of the central nervous system that causes difficulty in movement, often including tremors and rigidity. Early detection of PD can prevent symptoms up to a certain age and increase life expectancy. For this purpose, we have used brain images from magnetic resonance imaging (MRI) technique. Generally dementia can be either classified as Alzheimer’s or Parkinson’s or sometimes may be due to tumor in brain. Therefore, effectual methods such as Competitive Swarm Coot Optimization_ Convolutional Neural Network (CSCOOT_CNN) with transfer learning and Fractional CSCOOT_ deep neuro-fuzzy network (FCSCOOT_DNFN are newly introduced for classification of brain diseases. At first, input images are acquired from particular datasets, and then input images are given to the pre-processing stage.  In a pre-processing module, median filter is utilized for the elimination of noises. Afterward, pre-processed image is then subjected to feature extraction in which CNN features are extracted. In the level of classification, the images are classified into Parkinson by DNFN that is trained utilizing the introduced FCSCOOT algorithm. Furthermore, the FCSCOOT algorithm is newly designed by combination of Fractional Calculus (FC) with CSCOOT algorithm.