End-to-End Learned Random Walker for Seeded Image Segmentation
This work addresses seeded segmentation for biomedical imaging, specifically neuron segmentation, and is incremental as it builds on the Random Walker algorithm with learned components.
The paper tackles seeded image segmentation by learning edge weights for the Random Walker algorithm using a convolutional neural network, achieving state-of-the-art results on the CREMI neuron segmentation challenge.
We present an end-to-end learned algorithm for seeded segmentation. Our method is based on the Random Walker algorithm, where we predict the edge weights of the underlying graph using a convolutional neural network. This can be interpreted as learning context-dependent diffusivities for a linear diffusion process. Besides calculating the exact gradient for optimizing these diffusivities, we also propose simplifications that sparsely sample the gradient and still yield competitive results. The proposed method achieves the currently best results on a seeded version of the CREMI neuron segmentation challenge.