Hansenclever de F. Bassani

Segmentation of clinical structures from images of the human pelvic area

The radiotherapy treatment planning requires the delineation of the therapy structures that will be submitted to the radiation beams. When executed manually, this delineation is a slow process and can result in human errors due to the amount of X-ray computed tomography (CT) images that are analyzed in each radiotherapy planning. This process needs precision, minimizing the radiation on healthy areas, close to the target tissues. A new system for automatic segmentation of images of clinical structures is proposed in this work. The algorithm is based on multi-region growing followed by the watershed transform. The main contributions are the method of seed pixel selection and predicate of the multi-region growing algorithm and the segmentation results achieved. The system was tested in 400 images and its efficiency was measured by two different statistical methods, correlation and the t-test. The clinical structures of interest are the rectum, bladder and seminal vesicles.

Watershed transform for automatic image segmentation of the human pelvic area

We propose a new system for automatic image segmentation of organs of the human pelvic area. The algorithm is based on multi-region growing followed by a watershed transform. The main contributions are the method of selecting seed pixels and predicating the multiregion growing algorithm, and the segmentation results achieved. The system was tested in 400 images and its efficiency was measured by two different statistical methods - correlation and the t-test. The clinical application of the proposed technique is radiotherapy treatment, which depends on the delineation of the treated area. The organs of interest are the rectum, bladder and seminal vesicles.

Dimension Selective Self-Organizing Maps for clustering high dimensional data

High dimensional datasets usually present several dimensions which are irrelevant for certain clusters while they are relevant to other clusters. These irrelevant dimensions bring difficulties to the traditional clustering algorithms, because the high discrepancies within them can make objects appear too different to be grouped in the same cluster. Subspace clustering algorithms have been proposed to address this issue. However, the problem remains an open challenge for datasets with noise and outliers. This article presents an approach for subspace and projected clustering based on Self-Organizing Maps (SOM), that is called Dimensional Selective Self-Organizing Map. DSSOM keeps the properties of SOM and it is able to find clusters and identify their relevant dimensions, simultaneously, during the self-organizing process. The results presented by DSSOM were promising when compared with state of art subspace clustering algorithms.

Surface Reconstruction Method Based on a Growing Self-Organizing Map

This work introduces a method that produces triangular mesh representation of a target object surface. The new surface reconstruction method is based on Growing Self-organizing Maps, which learns both the geometry and the topology of the input data set. Each map grows incrementally producing meshes of different resolutions, according to different application needs. Experimental results show that the proposed method can produce triangular meshes having approximately equilateral faces, that approximate very well the shape of an object, including its concave regions and holes, if any.

Learning vector quantization with local adaptive weighting for relevance determination in Genome-Wide association studies

In Genome-Wide Association Studies (GWAS) huge amounts of genetic information are analyzed in order to discover how the observed variations, more specifically, the Single Nucleotide Polymorphisms (SNPs), are related with a certain trait of interest, such as the susceptibility for a disease. However, the high dimensionality observed in the datasets imposes significant challenges for methods that try to identify the relevant SNPs and their interactions. In particular, we emphasize the challenges imposed by the great amount of irrelevant dimensions shadowing information which is object of study. In this work, we present a prototype-based classification method, derived from Learning Vector Quantization (LVQ), in which the relevance of each input dimension is learned independently for each prototype. We validate our method in simulated datasets of GWAS with a significant number of dimensions (20, 50, or 100) in which few of them (from 2 to 5) are relevant. Such dimensions have to be identified. The proposed method presented promising results, showing graceful degradation when the number of irrelevant dimensions increases, in comparison with Multifactor Dimensionality Reduction (MDR), Generalized Relevance Learning Vector Quantization (GRLVQ) and Supervised Relevance Neural Gas (SRNG).

Online incremental supervised growing neural gas

Online learning algorithms are intrinsically designed to deal with large amounts of data because of the one-instance-at-a-time approach to the learning process, circumventing memory issues and enabling real time learning. However, most online algorithms require previous knowledge of the problem to predetermine the number of categories to be learned, or some other kind of meta-information that is not likely to be available to a generic system. In this work, an online, incremental algorithm, oiSGNG, is proposed, whose main features are: zero nodes initialization and the original batch SGNG node insertion mechanism [10]. The results improved on the state of the art in 5 out of 12 multiclass datasets.

Evolutionary Algorithm for 3D Object Reconstruction from Images

This work presents an evolutionary algorithm to reconstruct 3D objects based on images of them. In the proposed evolutionary algorithm, we describe a way to evolve 3D initial models to a target object by means of comparisons between images generated from the models and images of target object, in which the acquisition position is known. We proposed a modification in the standard evolutionary strategy algorithm that produced better and faster results in the class of problems at hand (i.e. when many genes must evolve mainly to a direction in the search space). Satisfactory results were achieved reconstructing simple 3D objects, such as an ellipsoid object and a pear from a sphere. Although performance decrease for more detailed objects such as a face, the proposed solution still converges.

MOEA/D with uniformly randomly adaptive weights

When working with decomposition-based algorithms, an appropriate set of weights might improve quality of the final solution. A set of uniformly distributed weights usually leads to well-distributed solutions on a Pareto front. However, there are two main difficulties with this approach. Firstly, it may fail depending on the problem geometry. Secondly, the population size becomes not flexible as the number of objectives increases. In this paper, we propose the MOEA/D with Uniformly Randomly Adaptive Weights (MOEA/D-URAW) which uses the Uniformly Randomly method as an approach to subproblems generation, allowing a flexible population size even when working with many objective problems. During the evolutionary process, MOEA/D-URAW adds and removes subproblems as a function of the sparsity level of the population. Moreover, instead of requiring assumptions about the Pareto front shape, our method adapts its weights to the shape of the problem during the evolutionary process. Experimental results using WFG41-48 problem classes, with different Pareto front shapes, shows that the present method presents better or equal results in 77.5% of the problems evaluated from 2 to 6 objectives when compared with state-of-the-art methods in the literature.

Sistema Autônomo de Reconhecimento e Contagem de Ovos de Aedes aegypti

The space-time monitoring of Aedes aegytpi population is very important important for the public health. The Aedes population is estimated from egg counting of traps named ovitrap. The task of counting, currently performed with microscope, is exhausting for anyone to perform, and it is subject to various types of errors. The Autonomous System for the Recognition and Egg Count SARCO, was developed to automatically perform such counts, more accurately and quickly. Artificial neural networks combined with techniques of image processing and statistics were applied to count the number of eggs in digitized images of ovitramp’s reeds. Image segmentation was applied using only color information, and the number of eggs was estimated from the image area occupied by eggs. Validation tests used a set of 18 pictures of ovitraps with different egg densities, collected from programs of dengue vector (Aedes-SMTP) in the cities of Recife, Ipojuca and Santa Cruz do Capibaribe. For this data set, SARCO obtained a counting error around 12 %, well below the 18 % error obtained by professionals counting with microscope. All results were statistically validated through hypothesis testing.

Dimension Selective Self-Organizing Maps With Time-Varying Structure for Subspace and Projected Clustering

Subspace clustering is the task of identifying clusters in subspaces of the input dimensions of a given dataset. Noisy data in certain attributes cause difficulties for traditional clustering algorithms, because the high discrepancies within them can make objects appear too different to be grouped in the same cluster. This requires methods specially designed for subspace clustering. This paper presents our second approach to subspace and projected clustering based on self-organizing maps (SOMs), which is a local adaptive receptive field dimension selective SOM. By introducing a time-variant topology, our method is an improvement in terms of clustering quality, computational cost, and parameterization. This enables the method to identify the correct number of clusters and their respective relevant dimensions, and thus it presents nearly perfect results in synthetic datasets and surpasses our previous method in most of the real-world datasets considered.