CellTypeGraph: A New Geometric Computer Vision Benchmark

TL;DR

CellTypeGraph introduces a novel benchmark for node classification in geo-referenced graphs, specifically applied to plant organ cells, facilitating the evaluation of geometric learning methods with precomputed features and a PyTorch data loader.

Contribution

The paper presents a new benchmark dataset for cell classification in plant organs, along with a comprehensive evaluation of recent graph neural network architectures.

Findings

DeeperGCN performs best among tested models.

Benchmark dataset is publicly available with precomputed features.

Facilitates development of geometric learning methods in biology.

Abstract

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. To allow the convenient testing of new geometrical learning methods, the benchmark of Arabidopsis thaliana ovules is made available as a PyTorch data loader, along with a large number of precomputed features. Finally, we benchmark eight recent graph neural network architectures, finding that DeeperGCN currently works best on this problem.