In this paper we propose C2G-Net, a pipeline for image classification that exploits the morphological properties of images containing a large number of similar objects like biological cells. C2G-Net consists of two components: (1) Cell2Grid, an image compression algorithm that identifies objects using segmentation and arranges them on a grid, and (2) DeepLNiNo, a CNN architecture with less than 10,000 trainable parameters aimed at facilitating model interpretability. To test the performance of C2G-Net we used multiplex immunohistochemistry images for predicting relapse risk in colon cancer. Compared to conventional CNN architectures trained on raw images, C2G-Net achieved similar prediction accuracy while training time was reduced by 85% and its model was is easier to interpret.
@article{arxiv.2007.03378,
title = {C2G-Net: Exploiting Morphological Properties for Image Classification},
author = {Laurin Herbsthofer and Barbara Prietl and Martina Tomberger and Thomas Pieber and Pablo López-García},
journal= {arXiv preprint arXiv:2007.03378},
year = {2020}
}
Comments
10 pages, 5 figures (Figure 3 with 4 sub-figures), Appendix A and Appendix B after the references. Originally submitted to ICML2020 but rejected