Silviu-Ioan Bejinariu

Parallel image registration using bio-inspired computing

In this paper it is proposed a parallel approach for the pixel intensity based image registration (IR) problem on multi-core processors. While IR is an optimization problem which computes the optimal parameters for a geometric transform, two classes of bio-inspired algorithms are studied: Bacterial Foraging Optimization Algorithm (BFOA) and Genetic Algorithm (GA). The optimal transform is applied to a source image in order to align it to a model image by maximizing a similarity measure. In the presented experiment, mutual information (MI) is used to evaluate the IR quality and most of the processing time is spent in this evaluation. The proposed parallel approach aims to reduce the processing time by using the full computing power of multi-core processors. A comparison of the sequential and parallel versions for different registration problems is presented.

Automatic multi-threshold image segmentation using metaheuristic algorithms

In this paper is presented an automatic segmentation approach for gray level images based on usage of metaheuristic swarming algorithms for multiple thresholds computing. The multi-threshold segmentation is an optimization problem while the thresholds must be determined and applied to the source image by minimizing an error measure. Because the number of possible solution may be very large in case of multiple thresholds, we used four metaheuristic swarming algorithms to obtain faster the optimal solution of the segmentation problem: Bacterial Foraging, Particle Swarming, Multi Swarm and Firefly optimization. As optimization criteria, root mean square error, peak signal-to-noise ratio and structural similarity index are used. Each optimization algorithm allows obtaining the optimal solution in a reasonable number of iterations and the obtained results were compared.

Image processing by means of some bio-inspired optimization algorithms

Image processing problems often require optimization algorithms to be applied. In this paper some aspects concerning the behavior of Bat and Cuckoo Search optimization algorithms are presented. The obtained accuracy and the processing time depend on the input images characteristics, chosen optimization criteria, dimension of the search space and, last but not least, on the chosen optimization algorithm and its parameters. The two nature inspired optimization algorithms were studied first in case of some mathematical functions minimization and then in case of bio-medical image registration.

A novel iris clustering approach using LAB color features

Interesting results of color clustering for the iris images in the UBIRISv1 database are presented. The iris colors are characterized by feature vectors with 80 components corresponding to histogram bins computed in the CIELAB color space. The feature extraction is applied to the first session eye images after undergoing an iris segmentation process. An agglomerative hierarchical algorithm is used to organize 1.205 segmented iris images in 8 clusters based on their color content.

Image registration using Bacterial Foraging Optimization Algorithm on multi-core processors

In this paper is proposed a parallel approach of the Bacterial Foraging Optimization Algorithm (BFOA) used for image registration (IR) on multi-core processors based systems. IR is an optimization problem computing optimal parameters of the transform to align a source image to a given model by maximizing a similarity measure. IR is an important step in image fusion: the fusion result is affected by the registration result quality. A relatively new registration method is based on the Bacterial Foraging Optimization Algorithm. Depending on the used optimization technique and similarity measure, the IR process may be time consuming. A comparison of the sequential and proposed parallel execution on multi-core systems, for pixel intensity and point feature based registration is presented.

Social behavior in bacterial foraging optimization algorithm for image registration

In this paper a strategy to enhance performances in Bacterial Foraging Optimization Algorithm (BFOA) based Image Registration (IR) is proposed. The goal of the IR optimization problem is to determine a geometric transform which maximizes the similarity to a model image when is applied to a source image. In each step of standard BFOA, the bacteria move in random directions. Using the social behavior of bacteria colony, the proposed method directs the movement towards bacteria with lower values of cost function. This method allows reducing the number of iterations required to obtain the optimal geometric transform and offers better approximations.

Fireworks Algorithm Based Image Registration

In the Image Processing (IP) domain, optimization algorithms have to be applied in many cases. Nature-inspired heuristics allow obtaining near optimal solutions using lower computing resources. In this paper the Fireworks Algorithm (FWA) behavior is studied for Image Registration (IR) problems. The IR results accuracy is analyzed for different types of images, mainly in case of pixel based registration using the Normalized Mutual Information. FWA is compared to Particle Swarming (PSO), Cuckoo Search (CSA) and Genetic Algorithms (GA) in terms of results accuracy and number of objective function evaluations required to obtain the optimal geometric transform parameters. Because the pixel based IR may fail in case of images containing graphic drawings, a features based IR approach is proposed for this class of images. Comparing to other nature inspired algorithms, FWA performances are close to those of PSO and CSA in terms of accuracy. Considering the required computing time, that is determined by the number of cost function evaluations, FWA is little slower than PSO and much faster than CSA and GA.

Nature-inspired algorithms based multispectral image fusion

Image Fusion is the combining process of relevant information from one, two or more images to create a single image which is more complete than any of the input ones. Image fusion is used in medical diagnosis in case of multi-modal images and also for multispectral images processing. Considering that the result of the image fusion process must maximize an evaluation measure, the fusion can be seen as an optimization procedure. In this paper, it is proposed an image fusion approach based on the usage of three nature-inspired optimization metaheuristics: Particle swarming, Cuckoo Search and Fireworks algorithms. As fusion technique, the weighted average in both spatial and transformed domain is used. The weights which maximize the fusion result evaluation measure are approximated using the nature-inspired algorithms. The proposed approach is applied for multispectral image fusion and the results obtained using the three optimization metaheuristics are compared.

New optic disc localization approach in retinal images

The paper proposes an optic disc recognition method in color retinal images. In a first step the optic disc area is identified using a quite complex methodology. Then the disc edges in the segmented area are extracted and a circular optic disc boundary approximation by a Hough transform is obtained.

Retinal Vessel Classification Technique

A retinal vessel classification procedure is proposed. From the image of thinned vessel network, landmarks are extracted and classified as branching, crossover and end points. Then a vascular graph is generated. Using a stratified graph edge labeling procedure the artery/vein map is built. In a first step the graph branches near the optic disc are localized and classified. Each label is propagated along the most significant segments linked to initial vessels. The next labeling phase aims the not processed branches starting from already classified vessels. Only branches and edges at crossings are labeled. Finally, using the current labels set, the uncertain cases are solved.