Roto-Translation Equivariant Convolutional Networks: Application to Histopathology Image Analysis
This work addresses the need for robust, rotation-invariant models in computational pathology, where tissue orientation should not affect predictions, offering a more reliable alternative to data augmentation methods.
The authors tackled the problem of achieving rotation invariance in convolutional networks for histopathology image analysis by proposing SE(2)-group convolution layers, which guarantee equivariance to rotations and translations, and showed consistent performance improvements on tasks like mitosis detection, nuclei segmentation, and tumor classification.
Rotation-invariance is a desired property of machine-learning models for medical image analysis and in particular for computational pathology applications. We propose a framework to encode the geometric structure of the special Euclidean motion group SE(2) in convolutional networks to yield translation and rotation equivariance via the introduction of SE(2)-group convolution layers. This structure enables models to learn feature representations with a discretized orientation dimension that guarantees that their outputs are invariant under a discrete set of rotations. Conventional approaches for rotation invariance rely mostly on data augmentation, but this does not guarantee the robustness of the output when the input is rotated. At that, trained conventional CNNs may require test-time rotation augmentation to reach their full capability. This study is focused on histopathology image analysis applications for which it is desirable that the arbitrary global orientation information of the imaged tissues is not captured by the machine learning models. The proposed framework is evaluated on three different histopathology image analysis tasks (mitosis detection, nuclei segmentation and tumor classification). We present a comparative analysis for each problem and show that consistent increase of performances can be achieved when using the proposed framework.