0000000000080034
AUTHOR
Kirsi Majava
Optimization-based techniques for image restoration
Kohinanpoisto on olennainen osa kuvankäsittelyä. FL Kirsi Majavan väitöskehittää uuden menetelmän, jolla paljon häiriötä sisältävästä kuvasta löytyvät sekä jyrkät rajat että sileät osat. Ne ovat perinteisen kuvankäsittelyn ongelma-alueita. Työssä kehitetään myös uusia, tehokkaita numeerisia ratkaisumenetelmiä erilaisille kohinanpoistotehtäville. Jos kuvia halutaan analysoida automaattisesti tietokoneella, on usein välttämätöntä aloittaa kuvankäsittely poistamalla kuvasta kohinaa ja muita vääristymiä. Jos kuvasta halutaan automaattisesti löytää ja tunnistaa siinä esiintyvät objektit, on lisäksi olennaista, että kohinanpoisto säilyttää mahdollisimman tarkasti objektien reunat. Usein kuvankäsi…
A Memetic Differential Evolution in Filter Design for Defect Detection in Paper Production
This article proposes a Memetic Differential Evolution (MDE) for designing digital filters which aim at detecting defects of the paper produced during an industrial process. The MDE is an adaptive evolutionary algorithm which combines the powerful explorative features of Differential Evolution (DE) with the exploitative features of two local searchers. The local searchers are adaptively activated by means of a novel control parameter which measures fitness diversity within the population. Numerical results show that the DE framework is efficient for the class of problems under study and employment of exploitative local searchers is helpful in supporting the DE explorative mechanism in avoid…
Robust refinement of initial prototypes for partitioning-based clustering algorithms
Non-uniqueness of solutions and sensitivity to erroneous data are common problems to large-scale data clustering tasks. In order to avoid poor quality of solutions with partitioning-based clustering methods, robust estimates (that are highly insensitive to erroneous data values) are needed and initial cluster prototypes should be determined properly. In this paper, a robust density estimation initialization method that exploits the spatial median estimate to the prototype update is presented. Besides being insensitive to noise and outliers, the new method is also computationally comparable with other traditional methods. The methods are compared by numerical experiments on a set of syntheti…
An enhanced memetic differential evolution in filter design for defect detection in paper production.
This article proposes an Enhanced Memetic Differential Evolution (EMDE) for designing digital filters which aim at detecting defects of the paper produced during an industrial process. Defect detection is handled by means of two Gabor filters and their design is performed by the EMDE. The EMDE is a novel adaptive evolutionary algorithm which combines the powerful explorative features of Differential Evolution with the exploitative features of three local search algorithms employing different pivot rules and neighborhood generating functions. These local search algorithms are the Hooke Jeeves Algorithm, a Stochastic Local Search, and Simulated Annealing. The local search algorithms are adap…