CellTypeGraph Benchmark

This repository contains the benchmark dataset proposed in: CellTypeGraph: A New Geometric Computer Vision Benchmark, Cerrone et al., CVPR2022 <strong>Paper Abstract:</strong> Classifying all cells in an organ is a relevant and difficult problem from plant developmental biology. We here abstract the problem into a new benchmark for node classification in a geo-referenced graph. Solving it requires learning the spatial layout of the organ including symmetries. <br> To allow the convenient testing of new geometrical learning methods, the benchmark of <em>Arabidopsis thaliana</em> ovules is made available as a PyTorch data loader, along with a large number of precomputed features. <strong>Raw Data Source:</strong> The benchmark raw data is derived from "A digital 3D reference atlas reveals cellular growth patterns shaping the <em>Arabidopsis</em> ovule, A. Vijayan et al., eLife 2021". <strong>Source code:</strong> Repository for using the benchmark: https://github.com/hci-unihd/celltype-graph-benchmark Repository for reproducing the experiments in the manuscript: https://github.com/hci-unihd/plant-celltype <strong>Index:</strong> Repository for using the benchmark (data-loaders, transforms, metrics): https://github.com/hci-unihd/celltype-graph-benchmark Repository for reproducing the experiments in the manuscript (experiments, training, GNN models, features extraction, visualization, inference): https://github.com/hci-unihd/plant-celltype <strong>Download integrity check (md5sum):</strong> raw_data.zip: fd0ecccdea684d156fd8ac182a22251b label_grs_surface.zip: 7aa603f3d3cddbf766679d388076269b