Research On Encoding And Matching Algorithm For Iris Image

With the development of information technology, the awareness of security issues isheightened that has led to a higher demand for secure and convenient individualidentification technology. However, traditional identification technology is inherentlyinsecure and cannot meet current requirement that leads to a massive rise in the interestfor biometric personal identification. Human iris is an extremely valuable source ofbiometric information, as it is a very complex structure unique to an individual.Compared with other biometric features, iris patterns are more stable and reliable.Furthermore, iris recognition systems can be non-invasive to their users. So the study ofiris identification is more and more attracting some people's attention. This articleproposes an improved algorithm for encoding iris texture image. Correlation coefficientsor Hamming Distances of codes for samples are calculated and the experimental resultsare analyzed.Firstly, the article discusses and summarizes the particular advantage of iris forrecognition. And every parts of iris recognition system are introduced, where extractingand matching features are keys. Then the preprocessing of iris image is analyzed,extracting and matching technology are mainly studied. Consult many recognitionalgorithms all over the world and analyze the characteristic of texture image. Thematching method based on invariable moment is applied to iris recognition for an attempt.But the experimental results demonstrate that the algorithm isn't suitable for irisrecognition.Iris images which have been preprocessed and normalized have various specialfeatures. The characteristic of amplitude and phase information of iris image is extractedfor iris recognition. Especially phase information is more important compared toamplitude information because the difference of amplitude information is not obvious andis easily influenced by outside factors such as contrast of iris image, intensity oflamp-house and amplification of optical system. One suitable model of thetwo-dimensional receptive field profiles encountered experimentally in cortical simplecells is the parameterized family of "2-DGabor filters". The texture of iris image with thecharacteristic of amplitude and phase information is extracted by the parameterizedfamily of "2-DGabor filters", which can capture properties of orientation selectivity,spatial frequency selectivity, and phase information of image by the greatest extent. Thepaper proposes an improved algorithm for encoding iris texture image based on 2-DGaborwavelet. The experimental results demonstrate that the proposed improved algorithm isfeasible for iris recognition. The algorithm of matching based on Hamming Distance or Correlation Coefficientare applied to matching two iris eigenvectors. The experimental results of 58 iris samplesare shown and analyzed with the theory of statistical decision, and two matchingalgorithm are compared.