Narrow Results

Due to noisy acquisition devices and variation in impression conditions, the ridgelines of fingerprint images are mostly corrupted by various kinds of noise causing cracks, scratches and bridges in the ridges as well as blurs. These cause matching errors in fingerprint recognition. For an effective recognition the correct ridge pattern is essential which requires the enhancement of fingerprint images. Segment by segment analysis of the fingerprint pattern yields various ridge direction and frequencies. By selecting a directional filter with correct filter parameters to match ridge features at each point, we can effectively enhance fingerprint ridges. This paper proposes a fingerprint image enhancement based on CNN Gabor-Type filters.

In this paper, we propose a hybrid computational geometry-gray scale algorithm that enhances fingerprint images greatly. The algorithm extracts the local minima points that are positioned on the ridges of a fingerprint, then, it generates a Delaunay triangulation using these points of interest. This triangulation along with the local orientations give an accurate distance and orientation-based ridge frequency. Finally, a tuned anisotropic filter is locally applied and the enhanced output fingerprint image is obtained. When the algorithm is applied to rejected fingerprint images from FVC2004 DB2 database by the veryfinger application, these images pass and experimental results show that we obtain a low false and missed minutiae rate with an almost uniform distribution over the database. Moreover, the application of the proposed algorithm enables the extraction of features from all low-quality fingerprint images where the equal error rate of verification is decreased from 6.50% to 5% using nondamaged low-quality images in the database.

This paper presents a novel approach to fingerprint singular point detection. Singular points (cores and deltas) are used for fingerprint classification, sub-classification and registration. This method exploits the stability of the directional field pattern extracted from singular point regions at different resolution levels. The procedure is invariant to translations, scaling and small rotations. A fingerprint sub-classification procedure was built based on the proposed singular point detection method. Two kinds of tests were conducted on a subset consisting of 955 NIST-14 fingerprint images. First, automatic and forensic expert sub-classifications were compared. Second, the consistency of the proposed method was measured comparing automatic sub-classification for two different rolls of the same fingerprint.

We present the design and implementation of a fingerprint-based remote authentication system. In this system, the fingerprint features captured at a client terminal are transmitted over a communication channel to a central server location where the verification takes place. In order to overcome replay attacks, we propose an original approach to fingerprint matching that relies on different fingerprint features for every execution of the authentication protocol. The paper also introduces an efficient approach for fingerprint image enhancement that constitutes the core of the feature extraction module included in the proposed authentication system. A public domain collection of fingerprint images is used to evaluate the system performance. The experimental results reveal that the proposed system can achieve good matching performance on this data collection and, furthermore, is able to accurately reject replay attacks.

Perspiration phenomenon is very significant to detect the liveness of a finger. However, it requires two consecutive fingerprints to notice perspiration, and therefore may not be suitable for real time authentications. Some other methods in the literature need extra hardware to detect liveness. To alleviate these problems, in this paper, to detect liveness a new texture-based method using only the first fingerprint is proposed. It is based on the observation that real and spoof fingerprints exhibit different texture characteristics. Textural measures based on gray level co-occurrence matrix (GLCM) are used to characterize fingerprint texture. This is based on structural, orientation, roughness, smoothness and regularity differences of diverse regions in a fingerprint image. Wavelet energy signature is also used to obtain texture details. Dimensionalities of feature sets are reduced by Sequential Forward Floating Selection (SFFS) method. GLCM texture features and wavelet energy signature are independently tested on three classifiers: neural network, support vector machine and K-nearest neighbor. Finally, two best classifiers are fused using the "Sum Rule''. Fingerprint database consisting of 185 real, 90 Fun-Doh and 150 Gummy fingerprints is created. Multiple combinations of materials are used to create casts and moulds of spoof fingerprints. Experimental results indicate that, the new liveness detection method is very promising, as it needs only one fingerprint and no extra hardware to detect vitality.

The extension of fluorescence measurements of samples from steady state to dynamic methods offers the possibility to extract valuable information at the micro and nano level. Imaging of fluorescent samples with nanosecond resolution often imposes challenging problems, especially when dealing with very weak optical and electrical signals. In this context, this paper discusses the detection of latent fingerprint samples with nanosecond resolution. Imaging of fingerprint samples, which are deposited on strongly fluorescing substrates, is carried out by the subsequent suppression of the unwanted background fluorescence emissions using the time-resolved optical technique. "signature" characterisation of fingerprint samples treated with fluorescent magnetic powders is also carried out.

We present an algorithm that compresses two-dimensional data, which are piece-wise smooth in one direction and have oscillatory events in the other direction. Fine texture, seismic, hyper-spectral and fingerprints have this mixed structure. The transform part of the compression process is an algorithm that combines the application of the wavelet transform in one direction with the local cosine transform (LCT) in the other direction. This is why it is called hybrid compression. The quantization and the entropy coding parts in the compression process were taken from SPIHT codec but it can also be taken from any multiresolution based codec such as EZW. To efficiently apply the SPIHT codec to a mixed coefficients array, reordering of the LCT coefficients takes place. When oscillating events are present in different directions as in fingerprints or when the image comprises of a fine texture, a 2D LCT with coefficients reordering is applied. These algorithms outperform algorithms that are solely based on the the application of 2D wavelet transforms to each direction with either SPIHT or EZW coding including JPEG2000 compression standard. The proposed algorithms retain fine oscillating events including texture even at a low bitrate. Its compression capabilities are also demonstrated on multimedia images that have a fine texture. The wavelet part in the mixed transform of the hybrid algorithm utilizes the Butterworth wavelet transforms library that outperforms the 9/7 biorthogonal wavelet transform.

A typical biometric system has three distinct phases. These are biometric data acquisition, feature extraction, and decision-making. The first step, the acquisition phase, is extremely important. If high quality images are not obtained, the next phase cannot operate reliably. Fingerprint recognition remains as one of the most prominent biometric identification methods. In this paper, we develop a prototype optical-based fingerprints data acquisition system using a CCD digital still camera to capture a complete impression of finger area required for accurately identifying an individual and present an image-based approach for online fingerprint recognition with the objective to increase the overall matching performance. The fingerprint images are matched based on features extracted with an adaptive learning vector quantization (LVQ) neural network to yield peak recognition of 98.6% for a random set of 300 test prints (100 fingers × 3 images). This system can be adopted as a multi-modal biometrics where two or more fingers are matched.

Fingerprints and source profiles of fine and coarse sands that originate from Central Inner Mongolia during Asian continental sandstorms (ACS) can be used to identify the origin of Asian sands and to trace them as they travel downwind. Soil samples collected at various land surfaces in Central Inner Mongolia were resuspended using a dry powder atomizer in an enclosure chamber. The resuspended sands were then sampled by two dichotomous samplers situated at the bottom of the enclosure chamber for fine (PM_2.5) and coarse (PM_2.5-10) sands, respectively. The chemical composition of sands, including water-soluble ionic species, metallic contents, and carbonaceous contents, were further analyzed. Results from resuspension tests indicated that the soils contained considerably more coarse particles than fine. Moreover, Mg, K, Al, and Fe in coarse sand had strong correlations with each other. The ratio of Mg, K, Fe (or Al) to Al (or Fe) and OC/EC in the coarse sands can be used as the fingerprints of Asian sands originating from Central Inner Mongolia.

The present paper is focused on the characterization of the fingerprints using directional morphological transformations. The interest of the proposed approaches to compute orientations field is not only useful in computing fingerprint orientation patterns but also for characterizing other structures containing anysotropies (for example, some microstructures in material images). Two approaches are investigated in this paper. The first one is a global approach based on the directional granulometries using line segments as structuring elements. Then, the notion of quatree structure is used to go from a global approach to a local one. The second method considers a local approach by using the concept of distance function. The distance function is computed by the supremum of directional erosions. It begins with a small structuring element by taking into account all orientations. Then the methodology increments the size of the structuring element until the structure is completely removed. The maximum of the distance function contains the sizes of the greater lines that can be included in the structure. To know the direction of the lines, a second image containing the orientations is built when the distance function is computed. Finally, both images, the distance function and the orientation function are used to estimate the lines at different orientations to characterize the fingerprint.