Title :
Identifying leaves with watershed and CSS: Implementation of segmentation and object matching
Author :
Su, Joey C. ; Po-Yuan Su ; Wei-Yang Lin
Author_Institution :
Dept. of Comput. Sci. & Inf. Eng., Nat. Chung Cheng Univ., Chiayi, Taiwan
Abstract :
This article looks at combining the Curvature Scale Space (CSS) shape matching algorithm and the Watershed Algorithm with the aim of creating a simple yet effective leaf identification software. The CSS shape matching algorithm solves the problems of: Leaf self-intersection - Parts of a leaf overlap. ; Leaf rotation - Images that are rotated. ; Leaf variation - No two leaves are the same. The Watershed Algorithm provides an effective way of segmenting a leaf from a non-ideal background, which provides a method of creating pristine binary images. The result is a leaf identification software that has good accuracy, the ability to deal with non-ideal backgrounds, and large room for improvement.
Keywords :
image colour analysis; image segmentation; object detection; CSS shape matching algorithm; binary images; curvature scale space shape matching algorithm; leaf identification software; leaf overlap; leaf rotation; leaf self-intersection; leaf variation; object matching; watershed algorithm; Cascading style sheets; Correlation; Image color analysis; Image segmentation; Shape; Software algorithms;
Conference_Titel :
Consumer Electronics - Taiwan (ICCE-TW), 2014 IEEE International Conference on
Conference_Location :
Taipei
DOI :
10.1109/ICCE-TW.2014.6904029