Deep Learning Models for Image Segmentation R package imageseg version 0.5.2 A general-purpose workflow for image segmentation using TensorFlow models based on the U-Net architecture by Ronneberger et al. (2015) <doi:10.48550/arXiv.1505.04597> and the U-Net++ architecture by Zhou et al. (2018) <doi:10.48550/arXiv.1807.10165>. We provide pre-trained models for assessing canopy density and understory vegetation density from vegetation photos. In addition, the package provides a workflow for easily creating model input and model architectures for general-purpose image segmentation based on grayscale or color images, both for binary and multi-class image segmentation. Software Imports: grDevices, keras, magick, magrittr, methods, purrr, stats, tibble, foreach, parallel, doParallel, dplyr Suggests: testthat Juergen Niedballa <camtrapr@gmail.com> Comprehensive R Archive Network (CRAN) Juergen Niedballa [aut, cre] (ORCID: <https://orcid.org/0000-0002-9187-2116>), Jan Axtner [aut] (ORCID: <https://orcid.org/0000-0003-1269-5586>), Leibniz Institute for Zoo and Wildlife Research [cph] MIT + file LICENSE (https://CRAN.R-project.org/package=imageseg/LICENSE) 2026-07-21 application/tgz https://CRAN.R-project.org/package=imageseg doi:10.32614/CRAN.package.imageseg