C2G-Net: Exploiting Morphological Properties for Image Classification
This work addresses image classification for medical applications like colon cancer relapse prediction, offering efficiency and interpretability gains, though it appears incremental as it builds on existing segmentation and CNN methods.
The paper tackled image classification for biological cells by proposing C2G-Net, a pipeline that uses morphological properties to compress images and a small CNN, achieving similar accuracy to conventional CNNs while reducing training time by 85% and improving interpretability.
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.