Dense Steerable Filter CNNs for Exploiting Rotational Symmetry in Histology Images

6 Apr 2020  ·  Simon Graham, David Epstein, Nasir Rajpoot ·

Histology images are inherently symmetric under rotation, where each orientation is equally as likely to appear. However, this rotational symmetry is not widely utilised as prior knowledge in modern Convolutional Neural Networks (CNNs), resulting in data hungry models that learn independent features at each orientation. Allowing CNNs to be rotation-equivariant removes the necessity to learn this set of transformations from the data and instead frees up model capacity, allowing more discriminative features to be learned. This reduction in the number of required parameters also reduces the risk of overfitting. In this paper, we propose Dense Steerable Filter CNNs (DSF-CNNs) that use group convolutions with multiple rotated copies of each filter in a densely connected framework. Each filter is defined as a linear combination of steerable basis filters, enabling exact rotation and decreasing the number of trainable parameters compared to standard filters. We also provide the first in-depth comparison of different rotation-equivariant CNNs for histology image analysis and demonstrate the advantage of encoding rotational symmetry into modern architectures. We show that DSF-CNNs achieve state-of-the-art performance, with significantly fewer parameters, when applied to three different tasks in the area of computational pathology: breast tumour classification, colon gland segmentation and multi-tissue nuclear segmentation.

PDF Abstract
Task Dataset Model Metric Name Metric Value Global Rank Result Benchmark
Colorectal Gland Segmentation: CRAG DSF-CNN (C8) F1-score 0.874 # 1
Dice 0.891 # 1
Hausdorff Distance (mm) 138.4 # 14
Multi-tissue Nucleus Segmentation Kumar DSF-CNN (C8) Dice 0.826 # 2
Hausdorff Distance (mm) 60 # 1
Breast Tumour Classification PCam DSF-CNN (C8) AUC 0.975 # 1


No methods listed for this paper. Add relevant methods here