Harald Ganster

Automated melanoma recognition

A system for the computerized analysis of images obtained from ELM has been developed to enhance the early recognition of malignant melanoma. As an initial step, the binary mask of the skin lesion is determined by several basic segmentation algorithms together with a fusion strategy. A set of features containing shape and radiometric features as well as local and global parameters is calculated to describe the malignancy of a lesion. Significant features are then selected from this set by application of statistical feature subset selection methods. The final kNN classification delivers a sensitivity of 87% with a specificity of 92%.

Active fusion—a new method applied to remote sensing image interpretation

Abstract Todays computer vision applications often have to deal with multiple, uncertain and incomplete visual information. In this paper, we introduce a new method, termed “ active fusion ”, which provides a common framework for active selection and combination of information from multiple sources in order to arrive at a reliable result at reasonable costs. The implementation of active fusion on the basis of probability theory, the Dempster-Shafer theory of evidence and fuzzy sets is discussed. In a sample experiment, active fusion using Bayesian networks is applied to agricultural field classification from multitemporal Landsat imagery. This experiment shows a significant reduction of the number of information sources required for a reliable decision.

A Robust Affine Matching Algorithm Using an Exponentially Decreasing Distance Function

We describe a robust method for spatial registration, which relies on the coarse correspondence of structures extracted from images, avoiding the establishment of point correspondences. These structures (tokens) are points, chains, polygons and regions at the level of intermediate symbolic representation (ISR). The algorithm recovers conformal transformations (4 affine parameters), so that 2-dimensional scenes as well as planar structures in 3D scenes can be handled. The affine transformation between two different tokensets is found by minimization of an exponentially decreasing distance function. As long as the tokensets are kept sparse, the method is very robust against a broad variety of common disturbances (e.g. incomplete segmentations, missing tokens, partial overlap). The performance of the algorithm is demonstrated using simple 2D shapes, medical, and remote sensing satellite images. The complexity of the algorithm is quadratic on the number of affine parameters.

Feature selection in melanoma recognition

Melanoma, one of the most aggressive types of cancer, can be healed, if recognized in early stages. In order to automate the early recognition of skin cancer; a system that analyses digital epiluminescence microscopic images is used. After segmentation, 33 features representing shape and radiometric properties are calculated. In the paper the quality of the features is evaluated by applying several feature selection methods. The results show that with each selection method the feature set can be reduced to dimension four with nearly no loss of information. Results with classification rates of up to 75% are achieved and relations between selected features and medical criteria are observed.

Affine Matching of Intermediate Symbolic Representations

Spatial registration of images, features and symbols is important for the comparison of these entities, as well as for the fusion of visual information gathered from diverse sources. We introduce a method for spatial registration, which relies on the coarse correspondence of structures extracted from the images, and which does not require the establishment of point correspondences. These structures (tokens) are points, chains, polygons and regions at the level of intermediate symbolic representation (ISR). The conformai (4 affine parameter) transformation is found by a combination of hierarchical hypothesize and test, stepwise refinement of parameters, and parameter variation avoiding local minima. Since belief values guide the probability that a certain token is selected for correspondence, and many-to-many correspondences are possible, the method is very robust against a broad variety of common disturbances like incomplete segmentations, missing tokens or partial overlap. This is demonstrated using synthetic test data as well as quite complicated multisource medical images. The establishment of this kind of spatial relationships between different ISR data sets is used as one module of a system for information fusion in image understanding.

Classifying Pigmented Skin Lesions with Machine Learning Methods

We use a data set of 1619 pigmented skin lesions images from three categories (common nevi, dysplastic nevi, and melanoma) to investigate the performance of four machine learning methods on the problem of classifying lesion images. The methods used were k-nearest neighbors, logistic regression, artificial neural networks, and support vector machines. The data sets were used to train the algorithms on the following tasks: to distinguish common nevi from dysplastic nevi and melanoma, and to distinguish melanoma from common nevi and dysplastic nevi. Receiver operating characteristic curves were used to summarize the performance of the models.