Semi-supervised Instance Segmentation with a Learned Shape Prior

TL;DR

This paper introduces a semi-supervised instance segmentation method using a learned shape prior via a variational autoencoder, achieving competitive results with minimal training data and outperforming some supervised models.

Contribution

It presents a novel semi-supervised framework that leverages a learned shape prior for instance segmentation, reducing the need for extensive annotated data.

Findings

Achieves results comparable to fully supervised methods with limited data.

Outperforms pre-trained supervised models on all tested datasets.

Requires only a few dozen shape patches, including synthetic shapes.

Abstract

To date, most instance segmentation approaches are based on supervised learning that requires a considerable amount of annotated object contours as training ground truth. Here, we propose a framework that searches for the target object based on a shape prior. The shape prior model is learned with a variational autoencoder that requires only a very limited amount of training data: In our experiments, a few dozens of object shape patches from the target dataset, as well as purely synthetic shapes, were sufficient to achieve results en par with supervised methods with full access to training data on two out of three cell segmentation datasets. Our method with a synthetic shape prior was superior to pre-trained supervised models with access to limited domain-specific training data on all three datasets. Since the learning of prior models requires shape patches, whether real or synthetic data, we call this framework semi-supervised learning.