Krzysztof Misztal

Detection of Elliptical Shapes via Cross-Entropy Clustering

The problem of finding elliptical shapes in an image will be considered. We discuss the new solution which uses cross-entropy clustering, providing the theoretical background of this approach. The proposed algorithm allows search for ellipses with predefined sizes and position in the space. Moreover, it works well in higher dimensions.

R Package CEC

Abstract Cross-Entropy Clustering (CEC) is a model-based clustering method which divides data into Gaussian-like clusters. The main advantage of CEC is that it combines the speed and simplicity of k-means with the ability of using various Gaussian models similarly to EM. Moreover, the method is capable of the automatic reduction of unnecessary clusters. In this paper we present the R Package CEC implementing CEC method.

Detection of Elongated Structures with Hierarchical Active Partitions and CEC-Based Image Representation

In this paper, a method of elongated structure detection is presented. In general, this is not a trivial task since standard image segmentation techniques require usually quite complex procedures to incorporate the information about the expected shape of the segments. The presented approach may be an interesting alternative for them. In its first phase, it changes the representation of the image. Instead of a set of pixels, the image is described by a set of ellipses representing fragments of the regions of similar color. This representation is obtained using cross-entropy clustering (CEC) method. The second phase analyses geometrical and spatial relationships between ellipses to select those of them that form an elongated structure within an acceptable range of its width. Both phases are elements of hierarchical active partition framework which iteratively collects semantic information about image content.

Deep learning approach to bacterial colony classification

In microbiology it is diagnostically useful to recognize various genera and species of bacteria. It can be achieved using computer-aided methods, which make the recognition processes more automatic and thus significantly reduce the time necessary for the classification. Moreover, in case of diagnostic uncertainty (the misleading similarity in shape or structure of bacterial cells), such methods can minimize the risk of incorrect recognition. In this article, we apply the state of the art method for texture analysis to classify genera and species of bacteria. This method uses deep Convolutional Neural Networks to obtain image descriptors, which are then encoded and classified with Support Vector Machine or Random Forest. To evaluate this approach and to make it comparable with other approaches, we provide a new dataset of images. DIBaS dataset (Digital Image of Bacterial Species) contains 660 images with 33 different genera and species of bacteria.

Pupil and Iris Detection Algorithm for Near-Infrared Capture Devices

In this paper a simple and robust solution for the pupil and iris detection is presented. The procedure is based on simple operations, such as erosion, dilation, binarization, flood filling and Sobel filter and, with proper implementation, is effective. The novelty of the approach is the use of distances of black points from nearest white points to estimate and then adjust the position of the center and the radius of the pupil which is also used for iris detection. The obtained results are promising, the pupil is extracted properly and all the information necessary for human identification and verification can be extracted from the found parts of the iris. The paper, being both review and research, contains also a state of the art in the described topic.

Weighted Approach to Projective Clustering

k-means is the basic method applied in many data clustering problems. As is known, its natural modification can be applied to projection clustering by changing the cost function from the squared-distance from the point to the squared distance from the affine subspace. However, to apply thus approach we need the beforehand knowledge of the dimension.

Mahalanobis Distance-Based Algorithm for Ellipse Growing in Iris Preprocessing

We introduce a new algorithm for ellipse recognition. The approach uses Mahalanobis distance and statistical and analytical properties of circular and elliptical objects. At first stage of the algorithm the starting configuration of initial ellipse is defined. Next we apply a condition which describes how much the shape is ellipse-like on the boundary points.

Iris Pattern Recognition with a New Mathematical Model to Its Rotation Detection

The work deals with the iris pattern recognition as one of the most popular automated biometric ways of individual identification. It is based on the acquired eye images in which we localize the region of interest – the iris. This extremely data-rich biometric identifier is stable throughout human life and well protected as internal part of the eye. Moreover, it is genetic independent, so that we can use it to identify or verify people among huge population. This chapter will present the human vision nature focusing on defects and diseases that change the surface information of the iris. Also will be shown the main stream and the historical background of mathematical research resulting in a new algorithm for automatic iris feature extraction. A special attention is paid to the method developed to detect the iris rotation for accurate success rate under different destructive problems and environmental conditions. The obtained results after using the new mathematical model have proved the algorithm high success rate in iris pattern recognition.

Cross Entropy Clustering Approach to Iris Segmentation for Biometrics Purpose

This work presents the step by step tutorial for how to use cross en- tropy clustering for the iris segmentation. We present the detailed construction of a suitable Gaussian model which best fits for in the case of iris images, and this is the novelty of the proposal approach. The obtained results are promising, both pupil and iris are extracted properly and all the information necessary for human identification and verification can be extracted from the found parts of the iris.

Subspaces clustering approach to lossy image compression

In this contribution lossy image compression based on subspaces clustering is considered. Given a PCA factorization of each cluster into subspaces and a maximal compression error, we show that the selection of those subspaces that provide the optimal lossy image compression is equivalent to the 0-1 Knapsack Problem. We present a theoretical and an experimental comparison between accurate and approximate algorithms for solving the 0-1 Knapsack problem in the case of lossy image compression.