Tariq M. Khan

Efficient hardware implementation strategy for local normalization of fingerprint images

Global techniques do not produce satisfying and definitive results for fingerprint image normalization due to the non-stationary nature of the image contents. Local normalization techniques are employed, which are a better alternative to deal with local image statistics. Conventional local normalization techniques involve pixelwise division by the local variance and thus have the potential to amplify unwanted noise structures, especially in low-activity background regions. To counter the background noise amplification, the research work presented here introduces a correction factor that, once multiplied with the output of the conventional normalization algorithm, will enhance only the feature region of the image while avoiding the background area entirely. In essence, its task is to provide the job of foreground segmentation. A modified local normalization has been proposed along with its efficient hardware structure. On the way to achieve real-time hardware implementation, certain important computationally efficient approximations are deployed. Test results show an improved speed for the hardware architecture while sustaining reasonable enhancement benchmarks.

Efficient Hardware Implementation For Fingerprint Image Enhancement Using Anisotropic Gaussian Filter

A real-time image filtering technique is proposed which could result in faster implementation for fingerprint image enhancement. One major hurdle associated with fingerprint filtering techniques is the expensive nature of their hardware implementations. To circumvent this, a modified anisotropic Gaussian filter is efficiently adopted in hardware by decomposing the filter into two orthogonal Gaussians and an oriented line Gaussian. An architecture is developed for dynamically controlling the orientation of the line Gaussian filter. To further improve the performance of the filter, the input image is homogenized by a local image normalization. In the proposed structure, for a middle-range reconfigurable FPGA, both parallel compute-intensive and real-time demands were achieved. We manage to efficiently speed up the image-processing time and improve the resource utilization of the FPGA. Test results show an improved speed for its hardware architecture while maintaining reasonable enhancement benchmarks.

A spatial domain scar removal strategy for fingerprint image enhancement

Fingerprints are the oldest and most widely used form of biometric identification. Many researchers have addressed the fingerprint classification problem and significant progress has been made in designing automatic fingerprint identification systems (AFIS) over the past two decades. However, some design factors such as lack of reliable minutia extraction algorithms, difficulty in quantitatively defining a reliable match between fingerprint images, poor image acquisition, low contrast images create bottlenecks in achieving the desired performance. Noticeable among them is the fact that digitally acquired fingerprint images are rarely of perfect quality to be used directly with AFIS; one important step is fingerprint enhancement. Conventional fingerprint enhancement methods, such as Gabor and anisotropic filters, do fill the holes and gaps in ridge lines but lack the necessary capability to tackle scar lines. For scar lines, an explicit filling process is proposed that is a mix of Fourier and spatial domain strategies. The proposed method is to make use of the Fourier domain directional field to trace an appropriate candidate for the scar pixels to be replaced with. The necessary components of the process are locating scars, estimating directional field, and the filling strategy. This process can act as front-end to the subsequent Gabor and anisotropic diffusion filtering. The simulation results for synthetic, as well as real fingerprints, show improved performance regarding better extraction of genuine minutia points. HighlightsWe Model a new proposed method to make use of the Fourier domain directional field to trace an appropriate candidate for the scar pixels to be replaced with.The necessary components of the process are locating scars, finding the directional field, and the filling strategy.The strategy relies on the fact that in these linear scars, the ridge/valley pattern is still intact across the scar region.Using this information, the scar boundary is filled with appropriate normal region pixels using the local orientation field.This process can act as front-end to the subsequent Gabor and anisotropic diffusion filtering.

Contrast normalization steps for increased sensitivity of a retinal image segmentation method

Retinal vessel segmentation plays a major role in the detection of many eye diseases, and it is required to implement an automated algorithm for analyzing the progress of eye diseases. A variety of automated segmentation methods have been presented but almost all studies to date showed weakness in their low sensitivity toward narrow low-contrast vessels. A new segmentation method is proposed to address the issue of low sensitivity, by including modules such as principal component analysis-based color-to-gray conversion, scale normalization factors for improved narrow vessel detection, anisotropic diffusion filtering with an adequate stopping rule, and edge pixel-based hysteresis threshold. The impact of these additional steps is assessed on publicly available databases like DRIVE and STARE. For the case of DRIVE database, the sensitivity is raised from 73 to 75%, while maintaining the accuracy of 96.5%, and found to provide evidence of improved sensitivity.

Automatic retinal vessel extraction algorithm based on contrast-sensitive schemes

Retinal vessel segmentation plays a key role in the detection of numerous eye diseases, and its reliable computerised implementation becomes important for automatic retinal disease screening systems. A large number of retinal vessel segmentation algorithms have been reported, primarily based on three main steps including making the background uniform, second-order Gaussian detector application and finally the region-grown binarization. Although these methods improve the accuracy levels, their sensitivity to low-contrast vessels still needs attention. In this paper, some contrast-sensitive approaches are discussed that once embedded in the conventional algorithm results in improved sensitivity for a given retinal vessel extraction technique. The impact of these add-on modules is assessed on publicly available databases like DRIVE and STARE and found to provide promising results.

Computerised approaches for the detection of diabetic retinopathy using retinal fundus images: a survey

AbstractEye-related disease such as diabetic retinopathy (DR) is a medical ailment in which the retina of the human eye is smashed because of damage to the tiny retinal blood vessels in the retina. Ophthalmologists identify DR based on various features such as the blood vessels, textures and pathologies. With the rapid development of methods of analysis of biomedical images and advanced computing techniques, image processing-based software for the detection of eye disease has been widely used as an important tool by ophthalmologists. In particular, computer vision-based methods are growing rapidly in the field of medical images analysis and are appropriate to advance ophthalmology. These tools depend entirely on visual analysis to identify abnormalities in Retinal Fundus images. During the past two decades, exciting improvement in the development of DR detection computerised systems has been observed. This paper reviews the development of analysing retinal images for the detection of DR in three aspects: automatic algorithms (classification or pixel to pixel methods), detection methods of pathologies from retinal fundus images, and extraction of blood vessels of retinal fundus image algorithms for the detection of DR. The paper presents a detailed explanation of each problem with respect to retinal images. The current techniques that are used to analyse retinal images and DR detection issues are also discussed in detail and recommendations are made for some future directions.

Automatic Retinal Vessel Extraction Algorithm

Retinal vessel segmentation plays a key role in the detection of numerous eye diseases, and its reliable computerised implementation becomes important for automatic retinal disease screening systems. A large number of retinal vessel segmentation algorithms have been reported, primarily based on three main steps including uniforming background, using the second-order Gaussian detector and applying binarization. These methods though improve the accuracy levels, their sensitivity to low-contrast in vessels still needs attention. In this paper, some contrast-sensitive approaches are discussed and embedded in the conventional algorithms, resulting in improved sensitivity for a given retinal vessel extraction technique. The proposed method gives good performance on both publicly databases with the accurate vessel extraction on STARE database. The proposed unsupervised method achieves the accuracy of 94.41%, much better than some existing unsupervised methods and comparable to some supervised methods. Its efficiency with different image conditions, together with its simplicity and fast operation, makes the blood vessel segmentation application suitable for retinal image computer analyses such as automated screening for early diabetic retinopathy detection.

Coherence enhancement diffusion using Multi-Scale DFB

Diffusion filtering techniques are mostly used to enhance the ridge structure of a noisy fingerprint image. In these filtering techniques the measurement of local orientation is needed. The diffusion tensor used in these techniques reflects the local image structure, as in a structure tensor same set of eigenvectors are used. To control the diffusion along the direction of high coherence special Eigenvalues are chosen. It works well in enhancing the ridges but, it takes orientation angles implicitly by using local image structure (derivatives). As we know that the derivatives have undesirable property of enhancing noise which makes the process of finding the correct orientation more difficult. This gives a further motivation for the improved orientation field calculated by some more reliable mean, which can overcome such difficulties. Therefore, in this work a Multi-Scale DDFB is used which adaptively change the local neighborhood size with the image local contrast and feature width. Experimental results show that the proposed algorithm is noise robust and is more suitable for feature localization as compare to other coherence enhancement diffusion algorithms.

Calibrating second-moment matrix for better shape adaptation with bias term from directional filter bank

Anisotropic diffusion is an important noise reduction process in many pattern recognition systems. The local orientations and shape are encoded in a descriptor, called second-moment matrix, to be used as the central part of the process. For a noisy image, it becomes hard to find the elements of the matrix with reasonable accuracy. Two layers of subsequent Gaussian smoothing are applied conventionally: One is referred to as noise scale and the other integration scale. It is still not guaranteed that a second-moment matrix is correctly aligned with underlying pattern flow directions. It is largely believed that the integration scale handles this orientation discrepancy. Therefore, some researchers suggested a space-dependent integration scale strategy. We propose here a calibration strategy using directional filterbank (DFB). The second-moment matrix is rotated to go along the proper ridge directions with an angle correction term from DFB. The experiments conducted show promise for the calibrated anisotropic diffusion process in terms of improved recognition rate over that of the un-calibrated process.

Role of Image Contrast Enhancement Technique for Ophthalmologist as Diagnostic Tool for Diabetic Retinopathy

Analysing the retinal colour fundus is a critical step before any proposed computerised automatic detection of eye disease, especially Diabetic Retinopathy (DR). The retinal colour fundus image contains noise and varying low contrast of the blood vessel against its surrounding background. It makes it difficult to analyse the proper order of the vessels network for detecting DR disease progress. The invasive method Fluorescein Angiogram Fundus (FFA) resolves these problems, but is not recommended due to an agent injection that leads to other side effects on the patients health, in the worst cases death. In this research work, we propose a new image enhancement method based on a morphological operation along with proposed threshold based stationary wavelet transform for retinal fundus images and Contrast Limited Adaptive Histogram Equalisation (CLAHE) for the vessel enhancement. The experimental results show much better results than the FFA images. Experimental results are based on three databases of retinal colour fundus images and FFA images. The performance is evaluated by measuring the contrast enhancement factor of retinal colour fundus images and FFA images. The results show that the proposed image enhancement method is superior to other non-invasive image enhancement methods as well as invasive methods, thus it will play an important role in imaging retinal blood vessels. An average contrast improvement factor of 5.63 on colour fundus images is achieved as well as 5.57 on FFA images. This significant contribution to the enhancement of the contrast of retinal colour fundus will be one primary tool to reduce the use of an invasive method.