Peekaboo - Where are the Objects? Structure Adjusting Superpixels

TL;DR

This paper introduces a dynamic superpixel segmentation method that adapts to image structure, improving boundary adherence and reducing undersegmentation while maintaining fast runtime.

Contribution

It extends SLIC by dynamically adjusting superpixel resolution based on image structure, enhancing segmentation quality without increasing computational cost.

Findings

Improved boundary adherence compared to state-of-the-art methods.

Reduced undersegmentation error.

Maintains linear runtime similar to standard SLIC.

Abstract

This paper addresses the search for a fast and meaningful image segmentation in the context of $k$-means clustering. The proposed method builds on a widely-used local version of Lloyd's algorithm, called Simple Linear Iterative Clustering (SLIC). We propose an algorithm which extends SLIC to dynamically adjust the local search, adopting superpixel resolution dynamically to structure existent in the image, and thus provides for more meaningful superpixels in the same linear runtime as standard SLIC. The proposed method is evaluated against state-of-the-art techniques and improved boundary adherence and undersegmentation error are observed, whilst still remaining among the fastest algorithms which are tested.