Our partner Carlos Fernandes has published an article on the latest number of SIG evolution, called A Camera Obscura for Ants. This article describes from the scientific point of view his Pherography method, that uses ant colony algorithms for creating curious and nice effects on photography. A bit like our Kohonants, but without the Kohonen part.
Last Friday I presented the paper “PicSOM – content-based image retrieval with self-organizing maps ” of Laaksonen et al. This authors uses T-SOMS (Tree-based SOMs) to classify images based in their content: using distinct measure types, like colour, sFFT (shape Fast Fourier Transform) and others several SOM maps are created (one per measure). The solution to how to give weight to that measures is simple: the user feedback. A set of images is presented to the user in a web-based application, so he can select the interested ones (example: Wynona Ryder’s face) to obtain the next set of images (an example can be seen in Figure 1). The algorithm learns with the selected images and gives weight to select the images of an specific map (in our example, shape measures are more important than colour measure).
Figure 1: Winona Ryder’s face in PicSom interface
To test the performance of every map and the global performance of the whole system the authors use sumatory things to stablish if the selected images belongs to a determined class (i.e. Planes, dinosaurs of faces). The conclusion is that it is necessary to use the whole set of measures at the same time to acquire the best performance. But the most interesting part is that you can read the paper and test the algorithm by yourself, a not so common practice.
Today, due to lack of volunteers, it was my turn again to deliver a paper seminar, and I chose Biogeography-Based Optimization, recently published in IEEE Transactions on Evolutionary Computation.
Interesting paper: a new nature-inspired algorithm from first principles. Conceptually, it is quite similar to estimation of distribution algorithms. The main concept is that of habitat, which have characteristics (represented by a vector) that make them suitable for a certain number of species; the better the HSI (habitat suitability index), the higher the number of species it can support. The population can migrate from one habitat to another, with flow going from those with more population to those with less population. Habitat characteristics can also change, in a way similar to mutation. Migration takes the place of crossover, with habitat variables migrating from those with a high population with those with a low population.
This more or less simple model yields rather good results, although it will probably be developed further in the future. It would be interesting, for instance, to add population structures, or to see how it could be paralellized.
GeNeura Team 