Automatic Individual Detection and Separation of Multiple Overlapped Nematode Worms Using Skeleton Analysis

Authors:
Nikzad Babaii Rizvandi;Aleksandra Pižurica;Wilfried Philips
Affiliations:
Image Processing and Interpretation Group (IPI), Department of Telecommunications and Information Processing (TELIN), Ghent University, Gent, Belgium 9000;Image Processing and Interpretation Group (IPI), Department of Telecommunications and Information Processing (TELIN), Ghent University, Gent, Belgium 9000;Image Processing and Interpretation Group (IPI), Department of Telecommunications and Information Processing (TELIN), Ghent University, Gent, Belgium 9000
Venue:
ICIAR '08 Proceedings of the 5th international conference on Image Analysis and Recognition
Year:
2008

Citing 5
Cited 1

Ziplock Snakes

International Journal of Computer Vision
Image Interpolation using Mathematical Morphology

DIAL '06 Proceedings of the Second International Conference on Document Image Analysis for Libraries
Computerized cell image analysis: past, present, and future

SCIA'03 Proceedings of the 13th Scandinavian conference on Image analysis
Detection of individual specimens in populations using contour energies

ACIVS'07 Proceedings of the 9th international conference on Advanced concepts for intelligent vision systems
Contour energy features for recognition of biological specimens in population images

ICIAR'07 Proceedings of the 4th international conference on Image Analysis and Recognition

Skeleton simplification by key points identification

MCPR'10 Proceedings of the 2nd Mexican conference on Pattern recognition: Advances in pattern recognition

Quantified Score

Hi-index	0.00

Visualization

Abstract

We present a new method for detection and separation of individual nematode worms in a still image. After pre-processing stage, which includes image binarization, filling the small holes, obtaining the skeleton of the new image and pruning the extra branches of skeleton, we split a skeleton into several branches by eliminating the connection pixels (pixels with more than 2 neighbors). Then we compute angles of all branches and compare the angles of the neighboring branches. The neighbor branches with angle differences less than a threshold are connected. Our method has been applied to a database of 54 overlap worms and results in 82% accuracy as automatic and 89% as semi-automatic with some limited user interaction.