EvoIASP is the first European event specifically dedicated to the applications of evolutionary computation (EC) to image analysis and signal processing (IASP) and gives European and non-European researchers in those fields, as well as people from industry, an opportunity to present their latest research and to discuss current developments and applications, besides fostering closer future interaction between members of the three scientific communities. All accepted papers will be presented orally at the conference and published by Springer as part of EvoWorkshops2004 in the Lecture Notes in Computer Science series.

LNCS 3005, the EvoWorkshops2004 proceedings, is now available online

EvoIASP PROGRAMME DETAILS

Workshop PAPERS:

Image segmentation by a Genetic Fuzzy c-Means Algorithm using color and spatial information
Lucia Ballerini, Leonardo Bocchi, Carina B. Johansson

Abstract:
This paper describes a new clustering algorithm for color image segmentation. We combine the classical fuzzy c-means algorithm (FCM) with a genetic algorithm (GA), and we modify the objective function of the FCM for taking into account the spatial information of image data and the intensity inhomogeneities. An application to medical images is presented. Experiments show that the proposed algorithm provides a useful method for image segmentation, without the need of a prefiltering step for background estimation. Moreover, the segmentation of noise images is effectively improved.

Genetic Optimization of Morphological Filters with Applications in Breast Cancer Detection
Lucia Ballerini, Lennart Franzen

Abstract:
In this paper we apply genetic algorithms to morphological filter optimization. The validation of the method is illustrated by performing experiments with synthetic images, whose optimal filter is known. Applications to microscopic images of breast tissue are reported. The medical problem consists in the discrimination between cancerous tissue and normal one.

Bond-Issuer Credit Rating with Grammatical Evolution
Anthony Brabazon, Michael O'Neill

Abstract:
This study examines the utility of Grammatical Evolution in modelling the corporate bond-issuer credit rating process, using information drawn from the financial statements of bond-issuing firms. Financial data, and the associated Standard & Poor's issuer-credit ratings of 791 US firms, drawn from the year 1999/2000 are used to train and test the model. The best developed model was found to be able to discriminate in-sample (out-of-sample) between investment-grade and junk bond ratings with an average accuracy of 87.59 (84.92)% across a five-fold cross validation. The results suggest that the two classifications of credit rating can be predicted with notable accuracy from a relatively limited subset of firm-specific financial data, using Grammatical Evolution.

Using GAs to Create a Waveguide Model of the Oral Vocal Tract
Crispin Cooper, David Howard, Andy M.Tyrrell

Abstract:
A Digital Waveguide Mesh is introduced as a method for acoustic modelling. Genetic Algorithms are applied to evolve the structure of a mesh to simulate the human throat and mouth, producing vowel-like sounds. The evolved shapes are compared to actual shapes adopted by the oral vocal tract during speech. The evolutionary methods are "tuned" for implementation on a new hardware device being developed under the POEtic project. Initial results given in this paper are promising.

Vision-Based Hand Motion Capture Using Genetic Algorithm
Jin-shi Cui, Zeng-qi Sun

Abstract:
3D hand motion capture plays an important role in multi-modal human-computer interfaces. Existing vision-based approaches mainly include two directions: model-based optimization framework and appearance-based learning approach. The main obstacle to handle with human hand motion capture is the high dimensionality associated with a full degree-of-freedom (DOF) articulated model. In this paper, a novel vision-based 3D hand motion capture algorithm is proposed. It views hand pose estimation and motion tracking as search prob-lems and utilizes genetic algorithm (GA). Firstly, a learning integrating with op-timization approach is introduced to estimate initial hand pose in 3D model based framework. And then a motion tracking method using GA-based particle filter (PF) is proposed to deal with the tracking problem in high-dimensional and multi-modal state space. Experimental results show that present approach significantly improves performance of motion tracking, especially in high-dimensional configuration space.

Top-Down Evolutionary Image Segmentation using a Hierarchical Social Metaheuristic
Abraham Duarte, Angel Sánchez, Felipe Fernández, Antonio S.Montemayor, Juan J. Pantrigo

Abstract:
This paper presents an application of a hierarchical social (HS) metaheuristic to region-based segmentation. The original image is modelled as a simplified image graph, which is successively partitioned into two regions, corresponding to the most significant components of the actual image, until a termination condition is met. The graph-partitioning task is solved as a variant of the min-cut problem (normalized cut) using an HS metaheuristic. The computational efficiency of the proposed algorithm for the normalized cut computation improves the performance of a standard genetic algorithm. We applied the HS approach to brightness segmentation on various synthetic and real images, with stimulating trade-off results between execution time and segmentation quality.

Multi-objective Sensor Planning for Efficient and Accurate Object Reconstruction
Enrique Dunn, Gustavo Olague

Abstract:
A novel approach for sensor planning, which incorporates multi-objective optimization principals into the autonomous design of sensing strategies, is presented. The study addresses planning the behavior of an automated 3D inspection system, consisting of a manipulator robot in an Eye-on-Hand configuration. Task planning in this context is stated as a constrained multi-objective optimization problem, where reconstruction accuracy and robot motion efficiency are the criteria to optimize. An approach based on evolutionary computation is developed and experimental results shown. The obtained convex Pareto front of solutions confirms the conflict among objectives in our planning.

An Algorithm for Segmenting Gaseous Objects on Images
Sung-Min Kim, Wonha Kim

Abstract:
A new methodology for segmenting gaseous object images is introduced. Unlike in case of a rigid object, the edge intensity of a gaseous object varies along the object boundary (edge intensities of some pixels on a gaseous object boundary are weaker than those of small rigid objects or noise itself). Therefore, the conventional edge detectors may not adequately detect boundary-like edge pixels of gaseous objects. We develop a novel object segmenting method using fuzzy algorithm trained by the genetic algorithm. The proposed method consists of a fuzzy-based boundary detector applicable to gaseous as well as rigid objects, and concave region filling to recover object regions. This algorithm is well applicable to medical image such as breast cancer or tumor segmentation.

Evolution Strategies approach for the solution of an inverse problem in Near-Field Optics
Demetrio Macias, Alexandre Vial, Dominique Barchiesi

Abstract:
We propose and study two inversion procedures for the characterization of a nanostructure from near-field scattered intensity data. For this approach, the inverse problem is reformulated as a non-linear constrained optimization problem. The solution of the resultant fitness function is found through the application of the $(\mu/\rho,\lambda)$ and $(\mu/\rho+\lambda)$ strategies. The performance of the inversion algorithms is illustrated through an example and the results are discussed by means of a comparison between the methods proposed.

A Watermark Sharing Scheme to High Quality Halftone Images with Genetic Algorithms
Emi Myodo, Kiyoshi Tanaka

Abstract:
In this paper, we propose a watermark sharing scheme to binary halftone images, which is extended from an image halftoning scheme using genetic algorithms. In addition to conventional evaluation functions on gray-level precision and appropriate contrast near edges, we design and introduce an evaluation function on watermark appearance to the scheme, and generate multiple binary halftone images shared a watermark image through solution search by GA. Also, we improve the resolution of the watermark image to the same size of generated images, which contributes to increase the appearance of watermark image remarkably. Simulation results show that we can share a watermark to generated halftone images keeping high image quality, but clearly decode the embedded watermark by overlapping those images optically.

Using Genetic Programming for Character Discrimination in Damaged Documents
Daniel Rivero, Juan R. Rabunal, Julián Dorado, Alejandro Pazos

Abstract:
In this paper we present an application of Genetic Programming (GP) to solve one problem in the field of image processing. This problem is the recovery of a deteriorated old document from the damages caused by centuries. This document was affected by many aggresive agents, mainly by the humidity caused by a wrong storage during many years. This makes this problem particularly hard and unaffordable by other image processing techniques. Recent works have shown how Genetic Algorithms is a technique suitable for this task, but in this paper we will show how we can obtain better results with GP.

Evolutionary Algorithm-Based Local Structure Modeling for Improved Active Shape Model
Jeongho Shin, Hyunjong Ki, Vivek Maik, Jinyoung Kang, Junghoon Jung, Joonki Paik

Abstract:
An evolutionary algorithm-based robust local structure modeling technique is proposed to improve the performance of the active shape model (ASM). The proposed algorithm can extract boundary of an object under adverse condition, such as noisy corruption, occlusions, and shadow effect. The principle idea of the evolutionary algorithm is to find the global minimum of an objective function by evolving from a large set of populations rather than a single solution which may cause a local minimum. The proposed algorithm has been tested for various images including a sequence of human motion to demonstrate the improved performance of object tracking based on the evolutionary ASM.

Program Size and Pixel Statistics in Genetic Programming for Object Detection
Mengjie Zhang, Urvesh Bhowan

Abstract:
This paper describes an approach to the use of genetic programming for object detection problems. In this approach, local region pixel statistics are used to form three terminal sets. The function set is constructed by the four standard arithmetic operators and a conditional operator. A multi-objective fitness function is constructed based on detection rate, false alarm rate, false alarm area and program size. This approach is applied to three object detection problems of increasing difficulty. The results suggest that the concentric circular pixel statistics are more effective than the square features for the coin detection problems. The fitness function with program size is more effective and more efficient for these object detection problems and the evolved genetic programs using this fitness function are much shorter and easier to interpret.

Intrinsic Evolvable Hardware in Digital Filter Design
Yang Zhang, Steve L. Smith, Andy M.Tyrrell

Abstract:
This paper presents the application of Intrinsic Evolvable Hardware to real-world combinational circuit synthesis, as an alternative to conventional approaches. The evolutionary technique employs Cartesian Genetic Program-ming at a functional level by devising compact evolutionary processing ele-ments and an external genetic reconfiguration unit. The experimental results conclude that in terms of computational effort, filtered image signal and im-plementation cost, the evolution outperforms convention approaches in most cases.

Multiclass Object Classification Using Genetic Programming
Mengjie Zhang, Will Smart

Abstract:
We describe an approach to the use of genetic programming for multiclass object classification problems. Rather than using fixed static thresholds as boundaries to distinguish between different classes, this approach introduces two methods of classification where the boundaries between different classes can be dynamically determined during the evolutionary process. The two methods are centred dynamic class boundary determination and slotted dynamic class boundary determination. The two methods are tested on four object classification problems of increasing difficulty and are compared with the commonly used static class boundary determination method. The results suggest that, while the static class boundary determination method works well on relatively easy object classification problems, the two dynamic class boundary determination methods outperform the static method for more difficult multiple class object classification problems.

Chair: Stefano Cagnoni, University of Parma <cagnoni@ce.unipr.it>
Giovanni Adorni (Italy) 
Lucia Ballerini (Sweden)
Bir Bhanu (USA)
Dario Bianchi (Italy)
Alberto Broggi (Italy)
Ela Claridge (UK)
Laura Dipietro (USA)
Marc Ebner (Germany)
Terry Fogarty (UK)
Daniel Howard (UK)
Mario Koeppen (Germany)
Evelyne Lutton (France)
Peter Nordin (Sweden)
Gustavo Olague (Mexico)
Riccardo Poli (UK)
Conor Ryan (Ireland)
Jim Smith (UK)
Giovanni Squillero (Italy)
Kiyoshi Tanaka (Japan)
Ankur M. Teredesai (USA)
Andy Tyrrell (UK)
Hans-Michael Voigt (Germany)
Mengjie Zhang (New Zealand)

Günther Raidl, Vienna University of Technology <raidl@ads.tuwien.ac.at>
Stefano Cagnoni, Universita' di Parma <cagnoni@ce.unipr.it>
Local chair : Ernesto Costa, University of Coimbra <ernesto@dei.uc.pt>

Evolutionary algorithms have been shown to be tools which can be used effectively in the development of systems (software or hardware) for image analysis and signal processing in complex domains of high industrial and social relevance.

EvoIASP is the first European event specifically dedicated to the applications of evolutionary computation (EC) to image analysis and signal processing (IASP) and gives European and non-European researchers in those fields, as well as people from industry, an opportunity to present their latest research and to discuss current developments and applications, besides fostering closer future interaction between members of the three scientific communities.

The previous editions of the Workshop were held in Göttenborg, Sweden (1999), Edinburgh, UK (2000), Como, Italy (2001), Kinsale, Ireland (2002), and Colchester, UK (2003).

The workshop is one of the activities of EvoIASP, the EvoNet working group on Evolutionary Computation for Image Analysis and Signal Processing.

Topics of interest include, but are not limited to:

• applications of evolutionary computation to real-life IASP problems,
• evolvable vision and signal processing hardware,
• evolutionary pattern recognition,
• hybrid architectures for machine vision and signal processing including evolutionary components,
• theoretical developments,
• comparisons between different evolutionary techniques and between evolutionary and non-evolutionary techniques in IASP applications,
• time series analysis by means of EC techniques.

Send your manuscript, at most 10 A4 pages long, in Springer LNCS format (instructions downloadable from http://www.springer.de/comp/lncs/authors.html) to the programme chair, Stefano Cagnoni, in gzipped PostScript or pdf format by email no later than 21 November (extended deadline) 2003. The papers will be peer reviewed by at least two members of the program committee. Authors will be notified via email on the results of the review by 19 December 2003.

The authors of accepted papers will have to improve their paper on the basis of the reviewers' comments and will be asked to send a camera ready version of their manuscripts, along with text sources and pictures, by 16 January 2003. The accepted papers will appear in the workshop proceedings, published in Springer LNCS Series, which will be available at the workshop.

The EvoWorkshops2004 proceedings will be
published by Spinger as part of their
Lecture Notes in Computer Science series.