Towards infield, live plant phenotyping using a reduced-parameter CNN

Atanbori, John; French, Andrew P.; Pridmore, Tony P.

Abstract

There is an increase in consumption of agricultural produce as a result of the rapidly growing human population, particularly in developing nations. This has triggered high-quality plant phenotyping re- search to help with the breeding of high yielding plants that can adapt to our continuously changing climate. Novel, low-cost, fully automated plant phenotyping systems, capable of in-field deployment are required to help identify quantitative plant phenotypes. The identification of quantitative plant phenotypes is a key challenge which relies heavily on the precise segmentation of plant images. Recently, the plant phenotyping community has started to use very deep Convolutional Neural Networks (CNNs) to help tackle this fundamental problem. How- ever, these very deep CNNs rely on some millions of model parameters and generate very large weight matrices, thus making them difficult to deploy in-field on low-cost, resource-limited devices.

Citation

Atanbori, J., French, A. P., & Pridmore, T. P. (2020). Towards infield, live plant phenotyping using a reduced-parameter CNN. Machine Vision and Applications, 31, Article 2. https://doi.org/10.1007/s00138-019-01051-7

• Journal Article Type: Article
• Acceptance Date: Nov 6, 2019
• Online Publication Date: Dec 17, 2019
• Publication Date: 2020-01
• Deposit Date: Sep 13, 2019
• Publicly Available Date: Sep 13, 2019
• Journal: Machine Vision and Applications
• Print ISSN: 0932-8092
• Electronic ISSN: 1432-1769
• Publisher: Springer Verlag
• Peer Reviewed: Peer Reviewed
• Volume: 31
• Article Number: 2
• DOI: https://doi.org/10.1007/s00138-019-01051-7
• Keywords: Hardware and Architecture; Software; Computer Vision and Pattern Recognition; Computer Science Applications
• Public URL: https://nottingham-repository.worktribe.com/output/2611579
• Publisher URL: https://link.springer.com/article/10.1007/s00138-019-01051-7
• Additional Information: Received: 2 October 2018; Revised: 11 April 2019; Accepted: 6 November 2019; First Online: 17 December 2019
• Contract Date: Sep 13, 2019