# Semi-Supervised Deep Learning for Fully Convolutional Networks

## TL;DR

This paper introduces a novel semi-supervised learning framework for Fully Convolutional Networks (FCNs), enhancing image segmentation performance by leveraging unlabeled data through auxiliary manifold embedding and Random Feature Embedding.

## Contribution

It presents the first semi-supervised learning method tailored for FCNs, applying auxiliary manifold embedding with Random Feature Embedding to improve segmentation tasks.

## Key findings

- Substantial performance improvements on MS Lesion Segmentation
- Effective domain adaptation using semi-supervised FCNs
- First semi-supervised approach for FCNs in image segmentation

## Abstract

Deep learning usually requires large amounts of labeled training data, but annotating data is costly and tedious. The framework of semi-supervised learning provides the means to use both labeled data and arbitrary amounts of unlabeled data for training. Recently, semi-supervised deep learning has been intensively studied for standard CNN architectures. However, Fully Convolutional Networks (FCNs) set the state-of-the-art for many image segmentation tasks. To the best of our knowledge, there is no existing semi-supervised learning method for such FCNs yet. We lift the concept of auxiliary manifold embedding for semi-supervised learning to FCNs with the help of Random Feature Embedding. In our experiments on the challenging task of MS Lesion Segmentation, we leverage the proposed framework for the purpose of domain adaptation and report substantial improvements over the baseline model.