Source Identification: A Self-Supervision Task for Dense Prediction

TL;DR

This paper introduces a novel self-supervision task called source identification for dense prediction, which improves feature learning in medical image segmentation by reconstructing original images from fused sources.

Contribution

The paper proposes a new self-supervision task inspired by blind source separation, demonstrating its effectiveness over traditional tasks in medical image segmentation.

Findings

SI task outperforms traditional self-supervision methods

Fusing images from different patients yields best results

Method improves dense prediction performance in medical imaging

Abstract

The paradigm of self-supervision focuses on representation learning from raw data without the need of labor-consuming annotations, which is the main bottleneck of current data-driven methods. Self-supervision tasks are often used to pre-train a neural network with a large amount of unlabeled data and extract generic features of the dataset. The learned model is likely to contain useful information which can be transferred to the downstream main task and improve performance compared to random parameter initialization. In this paper, we propose a new self-supervision task called source identification (SI), which is inspired by the classic blind source separation problem. Synthetic images are generated by fusing multiple source images and the network's task is to reconstruct the original images, given the fused images. A proper understanding of the image content is required to successfully solve the task. We validate our method on two medical image segmentation tasks: brain tumor segmentation and white matter hyperintensities segmentation. The results show that the proposed SI task outperforms traditional self-supervision tasks for dense predictions including inpainting, pixel shuffling, intensity shift, and super-resolution. Among variations of the SI task fusing images of different types, fusing images from different patients performs best.