Automatic image segmentation and analysis with applications to diatom identification.

In this thesis we have discussed automating the procedures for segmentation and identification of diatoms by digital image analysis. The main development in the first part of the thesis has been a framework for automatic segmentation of high magnification, grey-scale diatom images, which improved initial segmentation results obtained within the ADIAC project (Fischer et al. 2002, du Buf and Bayer 2002). The ultimate goal of the methods described in the second part is to perform automatic identification of diatoms, for which two multi-scale methods with roots in mathematical morphology were developed and employed for feature extraction and analysis of the segmented diatom images.

Zie: Summary

Samenvatting

Meer informatie in de catalogus
Meer informatie in Picarta