Computer Vision Problems in Plant Phenotyping (CVPPP) 2020


This year’s CVPPP was held alongside the European Conference on Computer Vision (ECCV) on Friday 28 August online.

All CVPPP papers, oral or poster, had a ten-minute pre-recorded video presentation which was made available to attendees, together with the paper, in advance of the workshop day. Workshop time focussed on interactive Q & A sessions, with each paper first being introduced by a one-minute teaser video.

Accepted papers addressed topics in detection, classification and scoring, segmentation, instance segmentation and counting, field tools and in-field operation, and time series and 3D. This year’s best paper prize was determined based on reviewer scores, rather than attendees’ votes, and was awarded to “Unsupervised Domain Adaptation For Plant Organ Counting” by Tewodros Ayalew, Jordan Ubbens and Ian Stavness. 

CVPPP has always incorporated technical challenges, and 2020 saw the introduction of the Global Wheat Detection Challenge. This was operated via Kaggle (https://www.kaggle.com/c/global-wheat-detection) and attracted a record 2200 submissions. The three best performing solutions were presented during the workshop.

Details of the CVPPP2020 programme are available at https://www.plant-phenotyping.org/CVPPP2020-Programme.  Full papers and videos are available via the ECCV workshop and tutorial site (login required).