TY - CONF
T1 - Segmentation, autofocusing and signature extraction of tuberculosis sputum images
PY - 2002/12/1
Y1 - 2002/12/1
N2 - Bacteria segmentation of particular species entails a challenging process. Bacteria shape is not enough as a discriminant feature, because there are many species that share the same shape. We present here two methods for tuberculosis image segmentation using the chromatic information. The first method is based on fuzzy segmentation of the color images based on the information that it is entailed in each separate chromatic histogram. The second method is a simple color filtering account by comparison of the inverse of the yellowish stained bacteria (blue channel) with the product of the other two chromatic channels. The third method is based on the extraction of image signatures by projecting logarithmic-polar mappings onto 1D vectors. This representation provides a very compact description of all image aspects, including shape, texture and color. An achromatic segmentation method is also presented based on the use of gray-level morphological operators only to the green channel. Finally we present the results of different autofocusing algorithms of stained tuberculosis images.
AB - Bacteria segmentation of particular species entails a challenging process. Bacteria shape is not enough as a discriminant feature, because there are many species that share the same shape. We present here two methods for tuberculosis image segmentation using the chromatic information. The first method is based on fuzzy segmentation of the color images based on the information that it is entailed in each separate chromatic histogram. The second method is a simple color filtering account by comparison of the inverse of the yellowish stained bacteria (blue channel) with the product of the other two chromatic channels. The third method is based on the extraction of image signatures by projecting logarithmic-polar mappings onto 1D vectors. This representation provides a very compact description of all image aspects, including shape, texture and color. An achromatic segmentation method is also presented based on the use of gray-level morphological operators only to the green channel. Finally we present the results of different autofocusing algorithms of stained tuberculosis images.
UR - https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=0036988486&origin=inward
UR - https://www.scopus.com/inward/citedby.uri?partnerID=HzOxMe3b&scp=0036988486&origin=inward
U2 - 10.1117/12.451665
DO - 10.1117/12.451665
M3 - Paper
SP - 171
EP - 182
T2 - Proceedings of SPIE - The International Society for Optical Engineering
Y2 - 1 December 2002
ER -