Unsupervised semantic discovery through visual patterns detection

TL;DR

This paper introduces a fast, fully unsupervised method for discovering semantic visual patterns that hierarchically categorize images and produce segmentation masks, outperforming previous methods in robustness and consistency.

Contribution

It presents a novel unsupervised framework with a new notion of semantic levels, a dedicated benchmark, and a two-phase algorithm for hierarchical visual pattern discovery.

Findings

Achieves high robustness to noise.

Produces semantically consistent segmentation masks.

Outperforms previous methods in experiments.

Abstract

We propose a new fast fully unsupervised method to discover semantic patterns. Our algorithm is able to hierarchically find visual categories and produce a segmentation mask where previous methods fail. Through the modeling of what is a visual pattern in an image, we introduce the notion of "semantic levels" and devise a conceptual framework along with measures and a dedicated benchmark dataset for future comparisons. Our algorithm is composed by two phases. A filtering phase, which selects semantical hotsposts by means of an accumulator space, then a clustering phase which propagates the semantic properties of the hotspots on a superpixels basis. We provide both qualitative and quantitative experimental validation, achieving optimal results in terms of robustness to noise and semantic consistency. We also made code and dataset publicly available.