Alcoholism Identification Based on an AlexNet Transfer Learning Model

Conclusion: This method can assist radiologists in their routine alcoholism screening of brain magnetic resonance images. Introduction Alcoholism (1) was previously composed of two types: alcohol abuse and alcohol dependence. According to current terminology, alcoholism differs from “harmful drinking” (2), which is an occasional pattern of drinking that contributes to increasing levels of alcohol-related ill-health. Today, it is defined depending on more than one of the following conditions: alcohol is strongly desired, usage results in social problems, drinking large amounts over a long time period, difficulty in reducing alcohol consumption, and usage resulting in non-fulfillment of everyday responsibilities. Alcoholism affects all parts of the body, but it particularly affects the brain. The size of gray matter and white matter of alcoholism subjects are less than age-matched controls (3), and this shrinkage can be observed using magnetic resonance imaging (MRI). However, neuroradiological diagnosis using MR images is a laborious process, and it is difficult to detect minor alterations in the brain of alcoholic patient. Therefore, development of a computer vision-based automatic smart alcoholism identification program is highly desirable to assist doctors in making a diagnosis. Within the last decade, studies have developed several promising alcoholism detection methods. Hou (4) put forward a novel algorithm called predator-prey adaptive-inertia c...
Source: Frontiers in Psychiatry - Category: Psychiatry Source Type: research