“Empyrean”, an alternative paradigm for pattern recognition

Infrared Thermal Mapping, Analysis and Interpretation in Biomedicine

“…The root of the tree is the entire scene, the leaves are the finest details and each region represents an object at a certain scale of observation [1]. Hierarchical Clustering-based Segmentation (HCS) [28][29][30] implements the traditional agglomerative clustering [23] where the regions of an initial partition are iteratively merged and automatically generate a hierarchy of segmented images (Fig. 5); for example, Fig.…”

Neural network based detection of hard exudates in retinal images

“…To select an adequate set of features, we focused on those characteristics that ophthalmologists use to visually distinguish EXs from the retinal background and other retinal lesions or structures. We also tried to keep an adequate dimensionality of the feature set, as misclassification probability tends to increase with the number of features and the structure of the classifier is more difficult to interpret [25].…”

Rough Sets and Rough Logic: A KDD Perspective

“…The choice of the language of formulas may be critical: on one hand, this language should be expressive enough to render all essential concepts in data structures. On the other hand too expressive a language may cause too high complexity of inference process (the phenomenon of language bias in Machine Learning, Pattern Recognition, KDD [40], [68], [73], [28]). The primitive data structures constructed according to a selected way(s) of recording a phenomenon present itself as possible models for various logics.…”

“…The problem of selecting relevant features [73], [70] involves some searching procedures like discretization, grouping of symbolic values, clusterization, morphological filtering. These preliminary procedures define primitive concepts (features) for a given problem of concept approximation.…”