Early Detection of Parkinson Disease By Using Dense Convolutional Neural Network Algorithm

Abstract

One of the most prevalent neurodegenerative illnesses affecting adults over 65 is Parkinson's disease. Due to the progressive nature of this illness, patients would suffer greatly from serious health-related issues as well as increased healthcare costs if this condition was not diagnosed in its early stages and monitored at various points. Parkinson's disease is a type of neurological illness that is typically found in older adults. Approximately 1% of people worldwide have Parkinson's disease (PD), and many of these patients have complex movement and cognitive problems. Cognitive and behavioural symptoms, such as a wide range of personality changes, depressive disorders, memory problems, and emotion dysregulation, may appear as the condition progresses. Furthermore, when the illness worsens, so do the symptoms related to movement. Early dementia diagnosis is important in order to apply suitable therapeutic strategies to stop cognitive decline. PD is often diagnosed by clinicians based on the presenting symptoms, which include stiffness, tremor, slowness, and issues with balance and coordination. However, each instance may have different symptoms and a different rate of advancement. As of right now, it appears that there is no particular blood test or biomarker to reliably identify Parkinson's disease (PD) or track underlying changes as the illness progresses. Magnetic resonance imaging, or MRI, has been utilized for the past thirty years to distinguish between different neurological conditions and probable Parkinson's disease. Numerous studies have demonstrated that current state-of-the-art CNNs maintain accuracy comparable to human ability. Furthermore, one of the most crucial elements in medical image processing has been feature representation. Conventional machine learning techniques would not take into account new and hidden features that are extracted and used by deep learning techniques like CNN. In order to evaluate the severity of the Parkinson's disease at different stages and between the healthy patient groups, an automated system that has been trained using features extracted from the various mri performed by both PD patients and the healthy group of patients has been developed in this project. The analysis was out in this study to separate the PD patients from the healthy volunteers based on the MRI by taking elements out of the drawings that the PD and healthy patients made of theit bodies.