Auto-thresholding Edge Detector for bio-image processing

Author

Yan Zhang ; Makowski, Lee

Author_Institution

Electr. & Comput. Eng., Northeastern Univ. Boston, Boston, MA, USA

fYear

2015

fDate

17-19 April 2015

Firstpage

1

Lastpage

2

Abstract

Canny Edge Detector has been widely used in biological related areas to detect edges of objects in an image. It has been recognized as an efficient edge detector, but the two thresholds that need to be provided by users remains an issue. In this paper, we develop an Auto-thresholding Edge Detector which applies k-means clustering algorithm for gradient intensity partitioning. The partitioned edge intensities are furthur utilized to compute high and low thresholds for the last step which is Hysteresis edge tracking. The Auto-thresholding Edge Detector solves the trial-and-error problem and doesn´t require users to input thresholds. The thresholds are calculated statistically based on individual images and the performance is promising.

Keywords

biological techniques; biology computing; edge detection; image processing; pattern clustering; statistical analysis; tracking; Canny edge detector; autothresholding edge detector; bioimage processing; edge intensity partitioning; gradient intensity partitioning; high threshold computation; hysteresis edge tracking; k-means clustering algorithm; low threshold computation; object edge detection; statistical calculation; threshold input; trial-and-error problem; Biology; Biomedical imaging; Clustering algorithms; Detectors; Hysteresis; Image edge detection; Partitioning algorithms;

fLanguage

English

Publisher

ieee

Conference_Titel

Biomedical Engineering Conference (NEBEC), 2015 41st Annual Northeast

Conference_Location

Troy, NY

Print_ISBN

978-1-4799-8358-2

Type

conf

DOI

10.1109/NEBEC.2015.7117209

Filename

7117209