Graph Cuts-Based Feature Extraction of Plant Leaf
As leaf is one of the most important organs in a plant, contour features of plant leaves are important for the identification of plant species. So researchers have proposed many methods to improve the progress of the plant identification. In this paper, we present a graph cuts-based method using Min-Cut/Max Flow algorithm to obtain the leaf blade section. Then, five basic features are computed to further obtain six digital morphological features. These experimental results show that the graph cuts algorithm and the presented leaf features are important for leaf recognition.
KeywordsPlant leaf Graph cuts Statistical features Feature extraction
The work reported in this paper was supported by Jinhua Polytechnic under the research grant 2011S002, and the Talent Start-up Foundation of Zhejiang A&F University under grant No. 2013FR059.
- 2.Daly DC, Hickey LJ, Johnson KR, Mitchell JD, Wilf P, Wing SL (2009) Manual of leaf architecture. CABI, AndersonGoogle Scholar
- 4.Ye Y, Chen C, Li CT, Fu H, Chi Z (2004) A computerized plant species recognition system. In: IEEE proceedings of 2004 international symposium on intelligent multimedia, video and speech processing, 2004 pp 723–726Google Scholar
- 6.Chien CF, Lin TT (2002) Leaf area measurement of selected vegetable seedlings using elliptical Hough transform. Trans ASAE 45(5):1669–1677Google Scholar
- 8.Eriksson AP, Barr O, Astrom K (2006) Image segmentation using minimal graph cuts. Published at: Swedish symposium on image. http://www.maths.lth.se/vision/publdb/reports/pdf/eriksson-barretal-ssia-06.pdf
- 13.Dinic EA (1970) Algorithm for solution of a problem of maximum flow in networks with power estimation. Soviet Math Dokl 11:1277–1280Google Scholar
- 17.Yuri B, Olga V, Ramin Z (2001) Efficient approximate energy minimization via graph Cuts. IEEE Trans PAMI 20(12):1222–1239Google Scholar