Oliveira Lab Publishes Deepweed: A Framework for Automated Individual Segmentation and Lineage Tracking of Duckweed

We’re excited to share that our paper, “Deepweed: Automated Individual Segmentation and Tracking for Duckweed Lineage Analysis,” is now published in Bioinformatics!

Duckweeds like Wolffia australiana are the smallest and fastest-growing flowering plants on Earth, with growing potential in biotechnology. They multiply through rapid clonal budding, yet most existing image analysis tools only measure growth at the population level, missing individual plants and the family trees connecting them.

Deepweed is a framework that fills that gap. Using deep-learning-based instance segmentation and multi-object tracking, it follows individual fronds across time-lapse images, assigns each one a persistent identity, and reconstructs mother-daughter relationships. The pipeline outputs lineage graphs, individual morphological metrics, and population growth summaries, along with lineage-annotated time-lapse videos. Applied to dense Wolffia populations, both free-growing and in microfluidic devices, Deepweed tracks hundreds of plants at once, resolves overlapping individuals, and rebuilds multi-generation lineages.

We are proud of our lab and especially of co-first authors Tolulope Oyeniran and Bukola Akindipe, who contributed equally to this work, under the supervision of PI Samuel M. D. Oliveira. We also want to thank our collaborators Tasmia Islam and Eric A. Josephs for bringing this challenge and dataset to us, and co-author Joshua D. Scheel for his contributions.

We can’t wait to help researchers across fields identify and analyze microscopy images of complex, crowded living systems.

Read the paper: https://doi.org/10.1093/bioinformatics/btag779

Citation: Oyeniran, T., Akindipe, B. A., Islam, T., Scheel, J. D., Josephs, E. A., & Oliveira, S. M. D. (2026). Deepweed: Automated individual segmentation and tracking for duckweed lineage analysis. Bioinformatics, btag779. https://doi.org/10.1093/bioinformatics/btag779