Martin Stommel

Learn More
It is shown that distance computations between SIFT-descriptors using the Euclidean distance suffer from the curse of dimensionality. The search for exact matches is less affected than the generalisation of image patterns, e.g. by clustering methods. Experimental results indicate that for the case of generalisation, the Hamming distance on binarised(More)
This paper describes a method to recognize and classify complex objects in digital images. To this end, a uniform representation of prototypes is introduced. The notion of a prototype describes a set of local features which allow to recognize objects by their appearance. During a training step a genetic algorithm is applied to the prototypes to optimize(More)
—Document enhancement tools are a valuable help in the study of historic documents. Given proper filter settings, many effects that impair the legibility can be evened out (e.g. washed out ink, stained and yellowed paper). However, because of differing authors, languages, handwritings, fonts and paper conditions, no single filter parameter set fits all(More)
In this paper we present and evaluate a simple but effective machine learning algorithm that we call Bitvector Machine: Feature vectors are partitioned along component-wise quantiles and converted into bitvectors that are learned. It is shown that the method is efficient in both training and classification. The effectiveness of the method is analysed(More)
— This paper addresses the problem of recognizing complex objects in images. The proposed approach is based on a prototype-centered object representation which describes objects as sets of local features. During an evolutionary learning step the model is derived from a set of sample images. The proceeding of the training is measured with regard to the(More)