Leaf segmentation and parallel phenotyping for the analysis of gene networks in plants

摘要: Over the last 4 years phenotyping is becoming more and automated, decreasing a lot of manual labour. Features, which uniquely define plant, can be extracted automatically from images. As plant data has to processed in order extract features, fast processing these features challenge. Therefore this paper, new method for automatic segmentation individual leaves plants with circular arrangement (rosettes) proposed, together an algorithm line symmetry leaf. Furthermore, achieve plants, four feature extraction methods are parallelised run on CPU GPU. Our evaluation results show that by parallelizing methods, it possible calculate image moments, area, histogram sum intensities 5 45 times faster than single threaded implementations.