Predicting Plant Growth from Time-Series Data Using Deep Learning

Phenotyping involves the quantitative assessment of anatomical, biochemical, and physiological plant traits. Natural growth cycles can be extremely slow, hindering experimental processes phenotyping. Deep learning offers a great deal support for automating addressing key phenotyping research issues. Machine learning-based high-throughput is potential solution to bottleneck, promising accelerate within phenomic research. This presents study deep networks’ predict plants’ expected growth, by generating segmentation masks root shoot systems into future. We adapt an existing generative adversarial predictive network this new domain. The results show efficient leaf that provides what system will look like at future time, based on time-series data growth. present benchmark two public datasets Arabidopsis (A. thaliana) Brassica rapa (Komatsuna) plants. strong performance, capability proposed methods match expert annotation. method highly adaptable, trainable (transfer learning/domain adaptation) different species mutations.