Video minor stroke extraction using learning vector quantization

Abstract

Video is very beneficial for health consultations, monitoring, disease identification, customer satisfaction survey efforts on fitness, health information in youtube. However, if the video health tips from youtube are untrue, the information will harm the person. Additionally, it will create the emotion youtube users who are looking for health information. Negative emotions cause effects blood pressure to rise. The impact of high blood pressure is such as stroke. In many developing countries found stroke, especially minor stroke disease which can attack several times a week. If it is not treated immediately, it will result in more severe disability. Handling stroke detection should be begun when the patient was thought to be affected. If it is done early, stroke treatment will reduce damage to organs caused. It requires early detection and treatment of stroke rehabilitation automatically. Therefore we propose the video object extraction minor stroke. The results of the segmentation process can be used to improve the detection performance of stroke in future studies. In addition, segmentation is used to maximize the prediction process of healing and rehabilitation of stroke patients in subsequent studies. This study uses a minor stroke extracted video object using LVQ which has been modified. Tests are using a variation of the constant (K) from 0.1 to 5. The results for the best accuracy are with a value 68.76 % of K = 4.3.