GANSeg: Learning to Segment by Unsupervised Hierarchical Image Generation

TL;DR

This paper introduces GANSeg, a novel unsupervised hierarchical image segmentation method that leverages mask-conditioned image generation to improve segmentation accuracy without requiring annotated masks.

Contribution

GANSeg is the first to use hierarchical mask conditioning in GANs for unsupervised image segmentation, reducing annotation needs and increasing robustness.

Findings

Outperforms state-of-the-art unsupervised segmentation methods

Generates high-quality image-mask pairs for training

Robust to viewpoint and object position changes

Abstract

Segmenting an image into its parts is a frequent preprocess for high-level vision tasks such as image editing. However, annotating masks for supervised training is expensive. Weakly-supervised and unsupervised methods exist, but they depend on the comparison of pairs of images, such as from multi-views, frames of videos, and image augmentation, which limits their applicability. To address this, we propose a GAN-based approach that generates images conditioned on latent masks, thereby alleviating full or weak annotations required in previous approaches. We show that such mask-conditioned image generation can be learned faithfully when conditioning the masks in a hierarchical manner on latent keypoints that define the position of parts explicitly. Without requiring supervision of masks or points, this strategy increases robustness to viewpoint and object positions changes. It also lets us generate image-mask pairs for training a segmentation network, which outperforms the state-of-the-art unsupervised segmentation methods on established benchmarks.