Automatic diatom identification using contour analysis by morphological curvature scale spaces

Abstract

A method for automatic identification of diatoms (single-celled algae with silica shells) based on extraction of features on the contour of the cells by multi-scale mathematical morphology is presented. After extracting the contour of the cell, it is smoothed adaptively, encoded using Freeman chain code, and converted into a curvature representation which is invariant under translation and scale change. A curvature scale space is built from these data, and the most important features are extracted from it by unsupervised cluster analysis. The resulting pattern vectors, which are also rotation-invariant, provide the input for automatic identification of diatoms by decision trees and k-nearest neighbour classifiers. The method is tested on two large sets of diatom images. The techniques used are applicable to other shapes besides diatoms.

A survey for the applications of content-based microscopic image analysis in microorganism classification domains

“…In Jalba et al (2005), a method is introduced to classify different diatom categories, where a multi-scale mathematical morphology feature is first extracted to represent the shape information, then the extracted features are used to train decision tree and k-NN classifiers. In the experiment, two data base are tested and obtain 75 and 90% accuracy, respectively.…”

Morphological hat-transform scale spaces and their use in texture classification

“…One way to solve this problem is to set the boundaries between classes of scale-space features from the data themselves. This is done by mean-shift cluster analysis as in [5]. After clustering is performed, the pattern vector is given by the centroids of the first six clusters with the largest areas.…”

“…In [5] we modified the initial technique of Leymarie and Levine to allow for nested structures, and included a method by which features in the scale space may be clustered in an unsupervised way, resulting in a small set of rotation, translation and scale-invariant shape parameters. In this paper we generalize the hat scale spaces to n-dimensional signals, give a fast algorithm for computing these scale spaces, and apply them to pattern classification.…”

“…However, the methods in [12] were tailored towards texture classification, while our method can also handle other types of classification problems, e.g. automatic identification of diatoms [5].…”

Time series adapted supervised fuzzy discretization: an application to ECG signals

“…Time series are datasets with continuous and quantitative characteristics that are encountered in real-world applications in areas like medicine, biology, and economics [1][2][3][4][5][6][7][8]. These continuous attributes are recorded as discrete signals by means of digital systems.…”