Humberto Sossa

A comparison of nature inspired algorithms for multi-threshold image segmentation

In the field of image analysis, segmentation is one of the most important preprocessing steps. One way to achieve segmentation is by mean of threshold selection, where each pixel that belongs to a determined class is labeled according to the selected threshold, giving as a result pixel groups that share visual characteristics in the image. Several methods have been proposed in order to solve threshold selection problems; in this work, it is used the method based on the mixture of Gaussian functions to approximate the 1D histogram of a gray level image and whose parameters are calculated using three nature inspired algorithms (Particle Swarm Optimization, Artificial Bee Colony Optimization and Differential Evolution). Each Gaussian function approximates the histogram, representing a pixel class and therefore a threshold point. Experimental results are shown, comparing in quantitative and qualitative fashion as well as the main advantages and drawbacks of each algorithm, applied to multi-threshold problem.

A multi-threshold segmentation approach based on Artificial Bee Colony optimization

This paper explores the use of the Artificial Bee Colony (ABC) algorithm to compute threshold selection for image segmentation. ABC is an evolutionary algorithm inspired by the intelligent behavior of honey-bees which has been successfully employed to solve complex optimization problems. In this approach, an image 1-D histogram is approximated through a Gaussian mixture model whose parameters are calculated by the ABC algorithm. In the model, each Gaussian function represents a pixel class and therefore a threshold point. Unlike the Expectation-Maximization (EM) algorithm, the ABC method shows fast convergence and low sensitivity to initial conditions. Remarkably, it also improves complex time-consuming computations commonly required by gradient-based methods. Experimental results over multiple images with different range of complexity validate the efficiency of the proposed technique with regard to segmentation accuracy, speed, and robustness. The paper also includes an experimental comparison to the EM and to one gradient-based method which ultimately demonstrates a better performance from the proposed algorithm.

Circle detection using electro-magnetism optimization

Nature-inspired computing has yielded remarkable applications of collective intelligence which considers simple elements for solving complex tasks by common interaction. On the other hand, automatic circle detection in digital images has been considered an important and complex task for the computer vision community that has devoted a tremendous amount of research, seeking for an optimal circle detector. This paper presents an algorithm for the automatic detection of circular shapes embedded into cluttered and noisy images without considering conventional Hough transform techniques. The approach is based on a nature-inspired technique known as the Electro-magnetism Optimization (EMO). It follows the electro-magnetism principle regarding a collective attraction-repulsion mechanism which manages particles towards an optimal solution. Each particle represents a solution by holding a charge which is related to the objective function to be optimized. The algorithm uses the encoding of three non-collinear points embedded into an edge-only image as candidate circles. Guided by the values of the objective function, the set of encoded candidate circles (charged particles) are evolved using an EMO algorithm so that they can fit into actual circular shapes over the edge-only map of the image. Experimental evidence from several tests on synthetic and natural images which provide a varying range of complexity validates the efficiency of our approach regarding accuracy, speed and robustness.

Adaptive and intelligent web based education system: Towards an integral architecture and framework

In this paper it is presented our contribution for carrying out adaptive and intelligent Web-based Education Systems (WBES) that take into account the individual student learning requirements, by means of a holistic architecture and Framework for developing WBES. In addition, three basic modules of the proposed WBES are outlined: an Authoring tool, a Semantic Web-based Evaluation, and a Cognitive Maps-based Student Model. As well, it is stated a Service Oriented Architecture (SOA) oriented to deploy reusable, accessible, durable and interoperable services. The approach enhances the Learning Technology Standard Architecture, proposed by IEEE-LTSA (Learning Technology System Architecture) [IEEE 1484.1/D9 LTSA (2001). Draft standard for learning technology - learning technology systems architecture (LTSA). New York, USA. URL: http://ieee.ltsc.org/wg1], and the Sharable Content Object Reusable Model (SCORM), claimed by Advanced Distributed Learning (ADL) [Advanced Distributed Learning Initiative (2004). URL: http://www.adlnet.org].

Design of artificial neural networks using a modified Particle Swarm Optimization algorithm

In the last years, bio-inspired algorithms have shown their power in different non-linear optimization problems. Due to the efficiency and adaptability of bio-inspired algorithms, in this paper we explore a new way to design an artificial neural network (ANN). For this task, a modified PSO algorithm was used. We do not only study the problem of finding the optimal synaptic weights of an ANN but also its topology and transfer functions. In other words, given a set of inputs and desired patterns, with the proposal we are able to find the best topology, the number of neurons, the transfer function for each neuron, as well as the synaptic weights. This allows to designing an ANN to be used to solve a given problem. The proposal is tested using several non-linear problems.

Retinal vessel extraction using Lattice Neural Networks with dendritic processing

Retinal images can be used to detect and follow up several important chronic diseases. The classification of retinal images requires an experienced ophthalmologist. This has been a bottleneck to implement routine screenings performed by general physicians. It has been proposed to create automated systems that can perform such task with little intervention from humans, with partial success. In this work, we report advances in such endeavor, by using a Lattice Neural Network with Dendritic Processing (LNNDP). We report results using several metrics, and compare against well known methods such as Support Vector Machines (SVM) and Multilayer Perceptrons (MLP). Our proposal shows better performance than other approaches reported in the literature. An additional advantage is that unlike those other tools, LNNDP requires no parameters, and it automatically constructs its structure to solve a particular problem. The proposed methodology requires four steps: (1) Pre-processing, (2) Feature computation, (3) Classification and (4) Post-processing. The Hotelling T(2) control chart was used to reduce the dimensionality of the feature vector, from 7 that were used before to 5 in this work. The experiments were run on images of DRIVE and STARE databases. The results show that on average, F1-Score is better in LNNDP, compared with SVM and MLP implementations. Same improvement is observed for MCC and the accuracy.

Efficient training for dendrite morphological neural networks

This paper introduces an efficient training algorithm for a dendrite morphological neural network (DMNN). Given p classes of patterns, C^k, k=1, 2, ..., p, the algorithm selects the patterns of all the classes and opens a hyper-cube HC^n (with n dimensions) with a size such that all the class elements remain inside HC^n. The size of HC^n can be chosen such that the border elements remain in some of the faces of HC^n, or can be chosen for a bigger size. This last selection allows the trained DMNN to be a very efficient classification machine in the presence of noise at the moment of testing, as we will see later. In a second step, the algorithm divides the HC^n into 2^n smaller hyper-cubes and verifies if each hyper-cube encloses patterns for only one class. If this is the case, the learning process is stopped and the DMNN is designed. If at least one hyper-cube HC^n encloses patterns of more than one class, then HC^n is divided into 2^n smaller hyper-cubes. The verification process is iteratively repeated onto each smaller hyper-cube until the stopping criterion is satisfied. At this moment the DMNN is designed. The algorithm was tested for benchmark problems and compare its performance against some reported algorithms, showing its superiority.

Artificial neural network synthesis by means of artificial bee colony (ABC) algorithm

Artificial bee colony (ABC) algorithm has been used in several optimization problems, including the optimization of synaptic weights from an Artificial Neural Network (ANN). However, this is not enough to generate a robust ANN. For that reason, some authors have proposed methodologies based on so-called metaheuristics that automatically allow designing an ANN, taking into account not only the optimization of the synaptic weights as well as the ANNs architecture, and the transfer function of each neuron. However, those methodologies do not generate a reduced design (synthesis) of the ANN. In this paper, we present an ABC based methodology, that maximizes its accuracy and minimizes the number of connections of an ANN by evolving at the same time the synaptic weights, the ANNs architecture and the transfer functions of each neuron. The methodology is tested with several pattern recognition problems.

Review: Causal knowledge and reasoning by cognitive maps: Pursuing a holistic approach

Due to the lack of an integral study about cognitive maps (CM) that focus on the causal phenomenon, this paper introduces the underlying concepts towards a holistic conceptual model, enhanced by a profile of several versions. We illustrate the use of CM through their application into the Web-based Education Systems (WBES). From the causal perspective, CM depict and simulate the systems dynamics based upon qualitative knowledge about a specific domain. A CM is a visual digraph that identifies the concepts of a given subject of analysis. CM show causal-effect relationships among the concepts and outline complex structures. This tool aims to predict the evolution of a model through causal inference. This kind of inference estimates the degree of significance of change of the concepts in the context of the whole system. The behavior of a CM is given away during iterations that update the variation of the concept state values until reach a stable point in a search space, a pattern of states or a chaotic region. The purpose of this research is to share its findings, depict the work done and promote the use of CM in a broad spectrum of domains.

Evolving ant colony system for optimizing path planning in mobile robots

Path planning is one of the problems in robotics. It consists on automatically determine a path from an initial position of the robot to its final position. In this paper we propose a variant of the ant colony system (ACO) applied to optimize the path that a robot can follow to reach its target destination. We also propose to evolve some parameters of the ACO algorithm by using a genetic algorithm (ACO-GA) to optimize the search of the shortest path. We compare the accuracy of ACO against ACO-GA using real environments.