Binary Morphological Neural Network
This work addresses a domain-specific problem for computer vision with binary images, but it is incremental as it adapts existing CNN concepts to mathematical morphology.
The authors tackled the problem of deep learning for binary images by creating a morphological neural network that replaces convolutions with erosions and dilations, resulting in promising experimental results for learning basic binary operators.
In the last ten years, Convolutional Neural Networks (CNNs) have formed the basis of deep-learning architectures for most computer vision tasks. However, they are not necessarily optimal. For example, mathematical morphology is known to be better suited to deal with binary images. In this work, we create a morphological neural network that handles binary inputs and outputs. We propose their construction inspired by CNNs to formulate layers adapted to such images by replacing convolutions with erosions and dilations. We give explainable theoretical results on whether or not the resulting learned networks are indeed morphological operators. We present promising experimental results designed to learn basic binary operators, and we have made our code publicly available online.