Dealing with Topological Information within a Fully Convolutional Neural Network
This addresses a specific problem in medical image segmentation for histological analysis, but appears incremental as it builds on existing methods.
The paper tackled the problem of fully convolutional neural networks lacking global topological information by proposing a solution based on pre-processing with a geodesic operator, applied to histological image segmentation of pigmented reconstructed epidermis via Whole Slide Imaging, but no concrete results or numbers were provided.
A fully convolutional neural network has a receptive field of limited size and therefore cannot exploit global information, such as topological information. A solution is proposed in this paper to solve this problem, based on pre-processing with a geodesic operator. It is applied to the segmentation of histological images of pigmented reconstructed epidermis acquired via Whole Slide Imaging.