Abstract

Multi-Modal image registration is the primary step in fusing complementary information contained in different imaging modalities for diagnostic purposes. We focus on two specific retinal imaging modalities namely, Color Fundus Image(CFI) and Fluroscein Fundus Angiogram(FFA). In this paper we investigate a projection based method using Radon transform for accurate matching in multi-modal retinal images. A novel Radon Transform based descriptor is proposed, which is invariant to illumination, rotation and partially to scale. Our results show that our descriptor is well suited for retinal images as it is robust to lesions, and works well even in poor quality images. The descriptor has been tested on a dataset of 126 images and compared for matching application against gradient based descriptors. The results show that Radon based descriptor outperforms the gradient based ones in both being able to discriminate between true and false matches and under presence of lesions.

Abstract

The presence of microaneurysms (MAs) is usually an early sign of diabetic retinopathy(DR) and their automatic detection from color retinal images is of clinical interest. In this paper, we present a new approach for automatic MA detection from digital colour fundus images. We formulate MA detection as a problem of target detection from clutter, where the probability of occurrence of target is considerably smaller compared to the clutter. A successive rejection-based strategy is proposed to progressively lower the number of clutter responses. The processing stages are designed to reject specific classes of clutter while passing majority of true MAs, using a set of specialized features. The true positives that remain after the final rejector are assigned a score which is based on its similarity to a true MA. Results of extensive evaluation of the proposed approach on three different retinal image datasets is reported, and are used to highlight the promise in the presented strategy.

Abstract

The aim of a local descriptor or a feature descriptor is to efficiently represent the region detected by an interest point operator in a compact format for use in various applications related to matching. The common design principle behind most of the mainstream descriptors like SIFT, GLOH, Shape context etc is to capture the spatial distribution of features using histograms computed over a grid around interest points. Histograms provide compact representation but typically loose the spatial distribution information. In this paper, we propose to use projection-based representation to improve a descriptor’s capacity to capture spatial distribution information while retaining the invariance required. Based on this proposal, two descriptors based on the CS-LBP are introduced. The descriptors have been evaluated against known descriptors on a standard dataset and found to outperform, in most cases, the existing descriptors. The obtained results demonstrate that proposed approach has the advantages of both the statistical robustness of histogram and the capability of the projection based representation to capture spatial information.

Abstract

We present a novel segmentation method to better capture the boundary of a non-homogeneous object such as the optic disk(OD), defined locally by two similar characteristic regions. Existing active contour models which utilise gradient information [12] or global region intensity [2] fail to localise aforementioned boundaries. We propose a region-based active contour model that uses local image information around each point of interest in multi-dimensional feature space. This model uses a local energy functional and level-set representation to achieve desired OD segmentation. The local energy functional defined on each image point provides sufficient information to determine a desired OD segmentation which is robust to the variations found in and around the OD region. This method has been evaluated against the segmentation provided by three medical experts on 138 retinal images. Both region and boundary-based assessment performed against two well established active contour models show strengths of the proposed method.

Abstract

Motion blur is one of the most common causes of image degradation. It is of increasing interest due to the deep penetration of digital cameras into consumer applications. In this paper, we start with a hypothesis that there is sufficient information within a blurred image and approach the deblurring problem as an optimisation process where the deblurring is to be done by satisfying a set of conditions. These conditions are derived from first principles underlying the degradation process assuming noise-free environments. We propose a novel but effective method for removing motion blur from a single blurred image via an iterative algorithm. The strength of this method is that it enables deblurring without resorting to estimation of the blur kernel or blur depth. The proposed iterative method has been tested on several images with different degrees of blur. The obtained results have been compared with state of the art techniques including those that require more than one input image. The results are consistently of high quality and comparable or superior to the existing methods which demonstrates the effectiveness of the proposed technique.

Abstract

In this paper, a novel and effective method for impulse noise removal in corrupted color images is discussed. The new method consists of two phases. The first phase is a noise detection phase where a modified Hopfield neural network is used to detect impulse noise pixels. The second is a noise filtering phase where the disadvantage of taking Vector Median in a single color space is addressed and a new algorithm based on performing Vector Median first in RGB space and then in HSI space is presented. The results of simulations performed on a set of standard test images on a wide range of noise corruption show that the proposed method is capable of detecting all the impulse noise pixels with almost zero false positive rates and removes noise while retaining finer image details. It outperforms the standard procedures and is yet simple and suitable for real time applications.

Abstract

We propose a method for synthetic zooming of tomographic images by applying super resolution technique on reconstructed data via a union of rotated lattices (URL). The proposed method consists of two steps: (i) sinogram data is filtered and back projected on to two lattices, which are rotated versions of each other and (ii) the samples from the two lattices are interpolated to generate the upsampled image. Square and hexagonal lattices have been investigated for URL. Results of subjective and objective evaluations of the proposed method on analytic phantoms are presented and compared with direct upsampling of data reconstructed on a single square lattice and upsampled image generated by union of low resolution shifted images (USL). The proposed method shows qualitative and quantitative improvement over direct up sampling but when compared with USL, generated up sampled images are of comparable quality.

Abstract

Registration is a primary step in tracking pathological changes in medical images. Point-based registration requires a set of distinct, identifiable and comparable landmark points to be extracted from images. In this work, we illustrate a method for obtaining landmarks based on changes in a topographic descriptor of a retinal image. Building on the curvature primal sketch introduced by Asada and Brady1 for describing interest points on planar curves, we extend the notion to grayscale images. We view an image as a topographic surface and propose to identify interest points on the surface using the surface curvature as a descriptor. This is illustrated by modeling retinal vessels as trenches and identifying landmarks as points where the trench behaviour changes, such as it splits or bends sharply. Based on this model, we present a method which uses the surface curvature to characterise landmark points on retinal vessels as points of high dispersion in the curvature orientation histogram computed around the points. This approach yields junction/crossover points of retinal vessels and provides a means to derive additional information about the type of junction. A scheme is developed for using such information and determining the correspondence between sets of landmarks from two images related by a rigid transformation. In this paper we present the details of the proposed approach and results of testing on images from public domain datasets. Results include comparison of landmark detection with two other methods, and results of correspondence derivation. Results show the approach to be successful and fast.

Abstract

There is a need to create quality open source educational resources to address the needs of Indian schools. In this paper,we describe a virtual physics lab which we built based on open source components. This is basically an authoring tool that lets user create their own experiments. We first discuss for the need for virtual labs. We then present the architecture of our tool and the technical challenges we faced in creating it. We also present the feedback we received from school students on this tool.

Abstract

Computed tomographic (CT) images are widely used in the diagnosis of stroke. In this paper, we present an automated method to detect and classify an abnormality into acute infarct, chronic infarct and hemorrhage at the slice level of non-contrast CT images. The proposed method consists of three main steps: image enhancement, detection of mid-line symmetry and classification of abnormal slices. A windowing operation is performed on the intensity distribution to enhance the region of interest. Domain knowledge about the anatomical structure of the skull and the brain is used to detect abnormalities in a rotation- and translation-invariant manner. A two-level classification scheme is used to detect abnormalities using features derived in the intensity and the wavelet domain. The proposed method has been evaluated on a dataset of 15 patients (347 image slices). The method gives 90% accuracy and 100% recall in detecting abnormality at patient level; and achieves an average precision of 91% and recall of 90% at the slice level.