Vision based Methodology for Diagnosis of Convulsion Patients

摘要

-Convulsion patients suffer from unforeseen loss of responsiveness likewise as uncontrolled jerking movements. It's additionally addressed as epileptic seizures. Most of the existing methods use electroencephalogram (EEG) for monitoring of epilepsy. The recording of encephalogram is made with the electrodes connected to scalp. This makes it effortful for the patient to stay with the attachment of the electrodes is effortful that makes long home monitoring that is byzantine. So as to manage these related problems, vision based methodology has been introduced in which video input is segmented first and segmented images are processed to detect & classify the type of convulsion. Optical flow calculation method is used for feature extraction and extracted features are distributed with the assistance of support vector machine. The introduced system is enabled with IoT that helps medical practitioner to look at the patient's status distantly. The major benefit of the introduced system is, it will hardly disrupt the regular sleep of the patient. This gives a price effective solution where patients can be monitored for long duration; thus it will contribute to enhance the diagnosis and so the quality of patient's life.