José M. Chaves-González

Detecting skin in face recognition systems: A colour spaces study

Skin colour detection is a technique very used in most of face detectors to find faces in images or videos. However, there is not a common opinion about which colour space is the best choice to do this task. Therefore, the motivation for our study is to discover which colour model is the best option to build an efficient face detector which can be embedded in a functional face recognition system. We have studied 10 of the most common and used colour spaces doing different comparisons among them, in order to know which one is the best option for human skin colour detection. In concrete, we have studied the models: RGB, CMY, YUV, YIQ, YPbPr, YCbCr, YCgCr, YDbDr, HSV-or HSI-and CIE-XYZ. To make the comparison among them, we have used 15 truth images where the skin colour of a face is clearly separated from the rest of the image (background, eyes, lips, hair, etc.). Thus we can compare at level pixel each colour model, doing a detailed study of each format. We present the final conclusions comparing different results, such as: right detections, false positives and false negatives for each colour space. According to the obtained results, the most appropriate colour spaces for skin colour detection are HSV model (the winner in our study), and the models YCgCr and YDbDr.

Metaheuristics for solving a real-world frequency assignment problem in GSM networks

The Frequency Assignment Problem (FAP) is one of the key issues in the design of GSM networks (Global System for Mobile communications), and will remain important in the foreseeable future. There are many versions of FAP, most of them benchmarking-like problems. We use a formulation of FAP, developed in published work, that focuses on aspects which are relevant for real-world GSM networks. In this paper, we have designed, adapted, and evaluated several types of metaheuristic for different time ranges. After a detailed statistical study, results indicate that these metaheuristics are very appropriate for this FAP. New interference results have been obtained, that significantly improve those published in previous research.

Optimization algorithms for large-scale real-world instances of the frequency assignment problem

Nowadays, mobile communications are experiencing a strong growth, being more and more indispensable. One of the key issues in the design of mobile networks is the frequency assignment problem (FAP). This problem is crucial at present and will remain important in the foreseeable future. Real-world instances of FAP typically involve very large networks, which can be handled only by heuristic methods. In the present work, we are interested in optimizing frequency assignments for problems described in a mathematical formalism that incorporates actual interference information, measured directly on the field, as is done in current GSM networks. To achieve this goal, a range of metaheuristics have been designed, adapted, and rigourously compared on two actual GSM networks modeled according to the latter formalism. To generate quickly and reliably high-quality solutions, all metaheuristics combine their global search capabilities with a local-search method specially tailored for this domain. The experiments and statistical tests show that in general, all metaheuristics are able to improve upon results published in previous studies, but two of the metaheuristics emerge as the best performers: a population-based algorithm (Scatter Search) and a trajectory based (1+1) Evolutionary Algorithm. Finally, the analysis of the frequency plans obtained offers insight about how the interference cost is reduced in the optimal plans.

A multiobjective swarm intelligence approach based on artificial bee colony for reliable DNA sequence design

The design of reliable DNA sequences is crucial in many engineering applications which depend on DNA-based technologies, such as nanotechnology or DNA computing. In these cases, two of the most important properties that must be controlled to obtain reliable sequences are self-assembly and self-complementary hybridization. These processes have to be restricted to avoid undesirable reactions, because in the specific case of DNA computing, undesirable reactions usually lead to incorrect computations. Therefore, it is important to design robust sets of sequences which provide efficient and reliable computations. The design of reliable DNA sequences involves heterogeneous and conflicting design criteria that do not fit traditional optimization methods. In this paper, DNA sequence design has been formulated as a multiobjective optimization problem and a novel multiobjective approach based on swarm intelligence has been proposed to solve it. Specifically, a multiobjective version of the Artificial Bee Colony metaheuristics (MO-ABC) is developed to tackle the problem. MO-ABC takes in consideration six different conflicting design criteria to generate reliable DNA sequences that can be used for bio-molecular computing. Moreover, in order to verify the effectiveness of the novel multiobjective proposal, formal comparisons with the well-known multiobjective standard NSGA-II (fast non-dominated sorting genetic algorithm) were performed. After a detailed study, results indicate that our artificial swarm intelligence approach obtains satisfactory reliable DNA sequences. Two multiobjective indicators were used in order to compare the developed algorithms: hypervolume and set coverage. Finally, other relevant works published in the literature were also studied to validate our results. To this respect the conclusion that can be drawn is that the novel approach proposed in this paper obtains very promising DNA sequences that significantly surpass other results previously published.

SS vs PBIL to solve a real-world frequency assignment problem in GSM networks

In this paper we study two different meta-heuristics to solve a real-word frequency assignment problem (FAP) in GSM networks. We have used a precise mathematical formulation in which the frequency plans are evaluated using accurate interference information coming from a real GSM network. We have developed an improved version of the scatter search (SS) algorithm in order to solve this problem. After accurately tuning this algorithm, it has been compared with a version fixed for the FAP problem of the population-based incremental learning (PBIL) algorithm. The results show that SS obtains better frequency plannings than PBIL for all the experiments performed.

Evolutionary algorithm based on different semantic similarity functions for synonym recognition in the biomedical domain

One of the most challenging problems in the semantic web field consists of computing the semantic similarity between different terms. The problem here is the lack of accurate domain-specific dictionaries, such as biomedical, financial or any other particular and dynamic field. In this article we propose a new approach which uses different existing semantic similarity methods to obtain precise results in the biomedical domain. Specifically, we have developed an evolutionary algorithm which uses information provided by different semantic similarity metrics. Our results have been validated against a variety of biomedical datasets and different collections of similarity functions. The proposed system provides very high quality results when compared against similarity ratings provided by human experts (in terms of Pearson correlation coefficient) surpassing the results of other relevant works previously published in the literature.

Teaching learning based optimization with Pareto tournament for the multiobjective software requirements selection

Software requirements selection is a problem which consists of choosing the set of new requirements which will be included in the next release of a software package. This NP-hard problem is an important issue involving several contradictory objectives which have to be tackled by software companies when developing new releases of software packages. Software projects have to stick to a budget, but they also have to satisfy the highest number of customer requirements. Furthermore, when managing real instances of the problem, the requirements tackled suffer interactions and other restrictions which make the problem even harder. In this paper, a novel multi-objective teaching learning based optimization (TLBO) algorithm has been successfully applied to several instances of the problem. For doing this, the software requirements selection problem has been formulated as a multiobjective optimization problem with two objectives: the total software development cost and the overall customer?s satisfaction. In addition, three interaction constraints have been also managed. In this context, the original TLBO algorithm has been adapted to solve real instances of the problem generated from data provided by experts. Numerical experiments with case studies on software requirements selection have been carried out in order to prove the effectiveness of the multiobjective proposal. In fact, the obtained results show that the developed algorithm performs better than other relevant algorithms previously published in the literature.

Differential evolution with Pareto tournament for the multi-objective next release problem

Software requirements selection is the engineering process in which the set of new requirements which will be included in the next release of a software product are chosen. This NP-hard problem is an important issue involving several contradictory objectives that have to be tackled by software companies when developing new releases of software packages. Software projects have to stick to a budget, but they also have to cover the highest number of customer requirements. Furthermore, in real instances of the problem, the requirements tackled suffer interactions and other restrictions which complicate the problem. In this paper, we use an adapted multi-objective version of the differential evolution (DE) evolutionary algorithm which has been successfully applied to several real instances of the problem. For doing this, the software requirements selection problem has been formulated as a multiobjective optimization problem with two objectives: the total software development cost and the overall customers satisfaction, and with three interaction constraints. On the other hand, the original DE algorithm has been adapted to solve real instances of the problem generated from data provided by experts. Numerical experiments with case studies on software requirements selection have been carried out to demonstrate the effectiveness of the multiobjective proposal and the obtained results show that the developed algorithm performs better than other relevant algorithms previously published in the literature under a set of public datasets.

DNA strand generation for DNA computing by using a multi-objective differential evolution algorithm

In this paper, we use an adapted multi-objective version of the differential evolution (DE) metaheuristics for the design and generation of reliable DNA libraries that can be used for computation. DNA sequence design is a very relevant task in many recent research fields, e.g. nanotechnology or DNA computing. Specifically, DNA computing is a new computational model which uses DNA molecules as information storage and their possible biological interactions as processing operators. Therefore, the possible reactions and interactions among molecules must be strictly controlled to prevent incorrect computations. The design of reliable DNA libraries for bio-molecular computing is an NP-hard combinatorial problem which involves many heterogeneous and conflicting design criteria. For this reason, we modelled DNA sequence design as a multiobjective optimization problem and we solved it by using an adapted multi-objective version of DE metaheuristics. Seven different bio-chemical design criteria have been simultaneously considered to obtain high quality DNA sequences which are suitable for molecular computing. Furthermore, we have developed the multiobjective standard fast non-dominated sorting genetic algorithm (NSGA-II) in order to perform a formal comparative study by using multi-objective indicators. Additionally, we have also compared our results with other relevant results published in the literature. We conclude that our proposal is a promising approach which is able to generate reliable real-world DNA sequences that significantly improve other DNA libraries previously published in the literature.

Software requirement optimization using a multiobjective swarm intelligence evolutionary algorithm

The selection of the new requirements which should be included in the development of the release of a software product is an important issue for software companies. This problem is known in the literature as the Next Release Problem (NRP). It is an NP-hard problem which simultaneously addresses two apparently contradictory objectives: the total cost of including the selected requirements in the next release of the software package, and the overall satisfaction of a set of customers who have different opinions about the priorities which should be given to the requirements, and also have different levels of importance within the company. Moreover, in the case of managing real instances of the problem, the proposed solutions have to satisfy certain interaction constraints which arise among some requirements. In this paper, the NRP is formulated as a multiobjective optimization problem with two objectives (cost and satisfaction) and three constraints (types of interactions). A multiobjective swarm intelligence metaheuristic is proposed to solve two real instances generated from data provided by experts. Analysis of the results showed that the proposed algorithm can efficiently generate high quality solutions. These were evaluated by comparing them with different proposals (in terms of multiobjective metrics). The results generated by the present approach surpass those generated in other relevant work in the literature (e.g. our technique can obtain a HV of over 60% for the most complex dataset managed, while the other approaches published cannot obtain an HV of more than 40% for the same dataset).