TY - JOUR AU - Dellen, B. AU - Scharr, Hanno AU - Torras, C. TI - Growth signature of rosette plants from time-lapse video JO - IEEE ACM transactions on computational biology and bioinformatics VL - PP IS - 99 SN - 1557-9964 CY - New York, NY PB - IEEE M1 - FZJ-2013-05546 SP - 1-11 PY - 2015 AB - Plant growth is a dynamic process, and the precisecourse of events during early plant development is of majorinterest for plant research. In this work, we investigate thegrowth of rosette plants by processing time-lapse videos ofgrowing plants, where we use Nicotiana tabacum (tobacco) asa model plant. In each frame of the video sequences, potentialleaves are detected using a leaf-shape model. These detectionsare prone to errors due to the complex shape of plants andtheir changing appearance in the image, depending on leafmovement, leaf growth, and illumination conditions. To copewith this problem, we employ a novel graph-based trackingalgorithm which can bridge gaps in the sequence by linkingleaf detections across a range of neighboring frames. We use theoverlap of fitted leaf models as a pairwise similarity measure, andforbid graph edges that would link leaf detections within a singleframe. We tested the method on a set of tobacco-plant growthsequences, and could track the first leaves of the plant, includingpartially or temporarily occluded ones, along complete sequences,demonstrating the applicability of the method to automatic plantgrowth analysis. All seedlings displayed approximately the samegrowth behavior, and a characteristic growth signature wasfound. LB - PUB:(DE-HGF)16 UR - <Go to ISI:>//WOS:000368292400027 DO - DOI:10.1109/TCBB.2015.2404810 UR - https://juser.fz-juelich.de/record/139562 ER -