Success in ImageCLEF Plant Identification Competition

Success in ImageCLEF Plant Identification Competition

 The Sabancı University-Okan University collaborative team won the first place in one of the two categories and third place in the other, in the ImageCLEF Plant Identification Competition.

 Plant recognition task consists of identifying the plant in a given image. This is a difficult task combining computer vision and pattern recognition fields, because of the high variability in the images in terms of (pose, scale, composition...) and high intra-class variations (due to plant maturity, season...) compared to high inter-class similarities.

 In the ImageCLEF Plant Identification competition organized in 2013, there were a total of 250 different species covering the majority of the plants in southern France. The photographs consisted of the entire plant or a leaf/flower/fruit/stem of the plant.

 Sabanci and Okan Universities participated in this competition as a team since 2011. The team combines their expertise in image processing and machine learning; the submitted system uses image descriptors based on mathematical morphology and signal processing and support vector machines as classifiers.

 In 2012, the team won the first place overall and in 2013 they won the first place in one category (uniform background) and third place in the other (natural background), among 12 international research teams and 33 submissions.

 Further details can be found at: http://www.imageclef.org/2013/plant