Metode Subtraksi Citra Sebagai Upaya Deteksi Gerakan Tangan

Abstract

Segmentation is one of the important operations in the object recognition process. One of the well-known segmentation methods is image subtraction. This image subtraction process is carried out to separate the foreground object from the background object. Most studies use this method to segment static images. In this study, the performance of the subtraction method will be observed based on the dynamic image input of hand movements. The application of the subtraction method is used for segmentation operations based on different information on different frames with the aim of obtaining information on the hand area. The focus of this research is to compare the results of processing with the subtraction method on different types of inputs based on the nature of their constituent signals, namely analog video input (webcam) and digital video input (stored video). The success of processing on both video inputs is determined from the first segmentation processing that finds the hand area. Two test cases (various lighting levels and outdoor background) were deliberately chosen because they are closely related to the potential weaknesses of the image subtraction method. Based on the test results, it is known that the implementation of the subtraction method for analog video input (webcam) and digital video input on various tests of light and background quality which is quite complex, shows success in finding areas of moving hands.

Keywords: Subtraction, Segmentation, Hand Movements