Machine vision
A fast parallel algorithm for thinning digital patterns
Communications of the ACM
Digital Image Processing
A general image based nematode identification system design
CIS'05 Proceedings of the 2005 international conference on Computational Intelligence and Security - Volume Part II
| Hi-index | 0.00 |
Nematodes are primitive creatures that are endangering and devouring many of the essential resources beneficial to human beings. For effective inspection and quarantine, we have devised an image based system for quantitatively characterizing and identifying nematodes, and achieved average successful identification rate of 71.2% for the Uncoordinated (Unc) mutant types and 91.2% for other types. To enhance system performance, here we introduce the worm-body Trunk Coordinate System for defining and characterizing the locomotion patterns of representative mutants. At least 60,000 frames for each species, totally 480,000 frames, representing wild type and 7 other mutant types, were analyzed. The average correct classification rate, measured by Classification and Regression Tree (CART) algorithm, was 79.3% for Unc types. The scheme devised and the features extracted are good supplements for the previous automated nematode identification system.