Search

SIFT is one of the feature region extraction methods. It is widely used in various fields including object recognition, mosaic object tracking, etc. However, since SIFT requires lots of arithmetic operations, it is not appropriate for real-time use. Therefore, many researchers has been tried to prepare other measures. SIFT using PCA or SURF is one of them. This paper is to study about the real-time CCTV combined system using the local partial image and to suggest the combination and recognition method for frames. Its purpose is to reduce the combination speed of the suggested algorithm compared to the existing one and to enhance the overall recognition rate according to this. Since the existing SIFT algorithm method has a patent and a slow processing speed, in this paper the improved local image regenerating method was used and the speed was raised according to the actual CCTV processing speed. Based on the recognition rate and speed by using the localized image, this paper consisted of realization of overall system and introduction of algorithm, and the comparative evaluation of performance was carried out through the actual experiment. By applying the CCTV operating in real time, the low-cost system was built with Inline System not with each image and this could reduce the physical fatigue of watcher, have an economic effect and could be used by general users easily.

Recently the importance of studies to recognize the shape of human hands in realtime is being more stressed to implement the user-friendly user interface. However, human hands have a large degree of freedom and thus it is very difficult to accurately recognize the hand shape, especially in a complicated background of the color same as the skin color. In this study, we detect the fingertip, without any special sensors, by using the hand shape based on the skin color information detected by a camera. This paper describes an algorithm to show directions using the detected fingertip. This algorithm is used to detect the fingertip through template matching and then detect forward, backward, left and right directions based on the detected fingertip and wrist center information. We applied the algorithm to navigation in 3D virtual-reality space, and achieved a very good result in an experiment on the change of directions to forward, backward, left and right. It is expected that this algorithm can provide more natural interaction between man and computer by allowing the user to instruct change of directions intuitively without the aid of a mouse, a keyboard or a joystick.