METHOD FOR DISCRIMINATING MALICIOUS VIDEO USING IMAGE QUALITY CLASSIFICATION AND APPARATUS THEREOF

摘要

Toxic video classification method to determine the image quality features of the present invention, at least one stop of the low-quality video generating an SVM (Support Vector Machine) learning model image and SVM learning speech model which is based on the sound characteristics based on the screen; Hazardous determined on the still image to extract the average of the input video GoF properties necessary by substituting for the SVM learning model image determining low quality; And with respect to the video in the video in low quality determined by extracting the audio information assigned to the SVM learning speech model determines whether the hazardous and toxic determining whether based on the color information with respect to a high-definition video determined; comprising and characterized in that, by applying an additional low-quality video and audio information with respect to possible to further improve the performance of existing hazardous video classification technique by determining whether the hazard, it is possible to develop a pest barrier through which the product marketable.