No abstract

In this paper, a correlation based fingerprint image segmentation technique is presented. Segmented is a useful image processing tool generally used in Machine vision applications in the Pattern Recognition and classification that leads to the separation of area of interest (AOI) from an image based on some image parameters like gray levels. In this proposed method, direction based segmentation is presented without computing and use of any directional field whose effectiveness has been evaluated for several finger print images (databases like FVC2002, FVC2000). KEYWORDSSegmentation, Fingerprint, Area of Interest (AOI), Correlation LITERATURE SURVEYThe foreground in a fingerprint image consists of oriented pattern with variations in orientations. In practice the presence of noise requires more robust segmentation techniques. Initially local or global thresholding techniques [3] used to segment foreground in a fingerprint image. Mehtre et al [4,5] isolated the fingerprint area based on the local histograms of ridge orientations. Ridge orientation is computed at each pixel and a histogram is computed for each 16 × 16 block. The presence of significant peaks in a histogram indicated an oriented pattern i.e. foreground, whereas a flat histogram characterize a background [4]. Mehtre et al proposed a composed method of segmentation in [5]. The local histograms of orientations and the gray-scale variance of each block have been used for segmentation. In the absence of reliable information from the histograms, assigns the lowvariance blocks to the background. This algorithm can handle the images where the background is uniform (white block in the background). Ratha et al [6] assigned each 16 × 16 block to the foreground or background according to the variance of gray-levels in an orthogonal direction to the ridge orientation. The background has no directional dependence and has uniform variance. The foreground has very high variance in a direction orthogonal to the ridge and very low variance along the ridges. Maio et al [7] separated the area of interest based on the average magnitude of the gradient in each image block. The fingerprint area is rich in edges due to ridges and valleys, the gradient response is high in the fingerprint area and small in the background. In the method proposed by Shen et al [8], eight Gabor filters are convolved with each image block, and the variance of the filter responses is used for fingerprint segmentation according to their quality, as "good," "poor," "smudged," or "dry." Bazen et al [9] proposed a pixel-wise segmentation technique. Where three features gradient coherence, intensity mean, and intensity variance are computed for each pixel and a linear classifier associates the pixel with the background or the foreground. A supervised technique is used to learn the optimal parameters for the linear classifier for each specific acquisition sensor. A final morphological post-processing step as suggested by Gonzalez et al [10] is performed to eliminate holes in both the foreground ...

No abstract

In this paper a novel method is presented for generation of a fixed length square finger code of size 4 16 16 × × . It uses a set of Gabor filters for extracting fingerprint features from gray scale image cropped in the size of 128 128 × pixels using its core point as the center. Experimental results show that the recognition rate based on the Euclidean distance between the two corresponding Gabor filter finger codes with verification accuracy of 93%. Since, the fingerprint matching is based on the Euclidean distance between two corresponding finger codes, it is extremely fast. This reveals that by setting the parameters to appropriate values, the finger code generated is more efficient and suitable than conventional methods for a small-scale fingerprint recognition system.

No abstract