영상관제에 있어 객체 식별률 향상을 위한 클래스 분류 및 이상신호 감지 모듈 개선에 관한 연구

Abstract

The field of security monitoring system and the various research has been conducted. However, anomalities detected in a video control system reached uppermost limit and system supervisor is not enough. Thus, this paper proposed that further improved intelligent monitoring system by tracking the path of movement of the object.. Currently, the following four important techniques are required for this task: MOG(Mixture-Of-Gaussian) and KNN(K-Nearest Neighbors), SIFT keypoint descriptor, method of least squares, there are some methods to obtain backgrounds and classifies class such as MOG and classifier KNN Algorithms builds a model using the training set., and then implementing the SIFT keypoint detector and descriptor extractor focusing on dynamic memory allocations. This trained classifier is expected to predict the label of new unseen samples. Conclusively, this paper improve KNN-Nearest Neighbors Algorithm for object classification by implementing The Scale Invariant Feature Transform (SIFT) keypoint detector and descriptor extractor focusing on dynamic memory allocations.