Improvement of texture based image segmentation algorithm for HE stained tissue samples

Windisch, Gergely; Kozlovszky, Miklós

doi:10.1109/cinti.2013.6705205

Cited by 3 publications

(2 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…It can be seen that the contours in the region do not match and none of the pixels should be considered a hit in the region however by pure chance the encircled pixel b elongs to contours on both images yielding a positive match according to (1). To tackle the issue of false positives we have devised a way derived fro m the Earth Movers Distance which is used many image processing applications to compare histograms [13].…”

Section: Framework For Comparison and Evaluation Of Image Segmentatiomentioning

confidence: 98%

See 1 more Smart Citation

Framework for Comparison and Evaluation of Image Segmentation Algorithms for Medical Imaging

Windisch

Kozlovszky

2015

IFMBE Proceedings

Self Cite

View full text Add to dashboard Cite

Abstract-Image segmentation algorithms are present in many image processing algorithms. Different algorithms achieve their goal differently, the usual method is to split the image into homogeneous segments based on some feature and then merge the neighboring segments that belong together. The present paper describes a method of comparison of image segmentation algorithms that focuses on the contour adherence accuracy of the given algorithms. A framework that has been designed and implemented is also presented here. The framework can and has been used for medical images (HE stained tissue digitalized with a microscope). The framework is presented using S LIC as a sample algorithm.

show abstract

Section: Framework For Comparison and Evaluation Of Image Segmentatiomentioning

confidence: 98%

“…Our research is in the field of medical imaging [1]. Our aim is to create supplementary services that can aid pathologists evaluate tissue samples.…”

Section: Introductionmentioning

confidence: 99%