Fractop: A Tool for Automated Biological Image Classification
David J. Cornforth, Herbert Franz Jelinek, Leo Peichl · Charles Sturt University Research Output (CRO) · 2002
Biological images often contain branching structures, especially those obtained from neural tissue.Neurons are known to fall into several types, but distinguishing these types is a continuing problem.Automatic classification of neuron type relies upon suitable measurements or features.Recently fractal dimension has been suggested as a useful feature.The fractal dimension is a measure of the complexity and selfsimilarity of an image, and is becoming accepted as a feature for automated classification of images having branching structures.Fractop is a web-based program designed to assist in the analysis of such images, but is applicable to any image.Here we describe a new version of this program, which is written in Java, enabling it to be used on a variety of platforms.Fractop is able to calculate the fractal dimension of an image using five variants of the mass radius technique.We present an example showing how Fractop, in conjunction with machine learning algorithms, can be used for automated classification of rat retinal ganglion cells.