FINGERPRINT RECOGNITION AND ID VERIFICATION USING WEBCAM

Abstract

While the use areas of biometric systems has not widen enough and the high costs as a result of this illustrate this technology as an expensive one. Because of this fact, in this research biometrical systems are examined and it's found meaningful to design a low-cost biometric system with keeping the use areas in mind. In this process, fingerprint recognition and ID verification using webcam concept is realized. Fingerprint verification is the method for recognition of the differences of lines and gaps over the human skin and comprasion of those. In this research this lines are transferred to MATLAB digital environment by using USB interface. By using adaptive treshnolding method and MATLAB's function image pre-processing as grey scaling and binarization, Median and FFT filtrering comparing, then for thinning MATLAB's algorithm, Least Square Method and Sobel processes are used in order to extract direction mapping. For minutiae marking, MATLAB's function was used and for individual fingerprint comparison false characterstics were removed from the taken fingerprint and mistakes were corrected. At last, successful results were obtained by using database, which was built by sixty fingerprint patterns of thirty different people, and user interface that was generated in MATLAB environment. It is observed that the success ratio of a sample which was gathered under suitable conditions, is %95.9184, false acception ratio is %3.33 and false rejection ratio is %1.66.