Perancangan sistem identifikasi biometrik jari tangan menggunakan Laplacian of Gaussian dan ektraksi kontur

Abstract

In this modern era, the identification system is becomes an important thing to be able to analyze a person based on the characteristics he has. In general, a person's identification system is still using conventional techniques, such as using ID cards that are easy to get off and lost passwords or PINs that are easily forgotten. In this research, the design of the system to identify or recognize a person using biometrics, especially the hand using Laplacian of Gaussian edge detection and Contour Extraction was carried out. The identification system with biometrics has many advantages because it is always attached to the body and can not be separated with humans. Stages performed in image processing in the form of RGB image conversion into binary using threshodling, edge detection using Laplacian of Gaussian as well as contour extraction using chain code representation. The result of contour extraction is the number of pixels calculated on each segment of the finger divided into 8 sections. The matching process on the system is done by using euclidean distance, ie calculate the smallest distance of similarity between the test image and the database. The result of identification conducted in this research using the feature extraction method, obtained accuracy of 93.33%.

Keywords: Finger Biometrics, Laplacian of Gaussian Edge Detection, Contour Extraction