SLIC Superpixels Compared to State-of-the-Art Superpixel Methods

TLDR

Superpixels are increasingly used in computer vision, yet defining a good superpixel algorithm remains unclear. The study compares five state‑of‑the‑art superpixel algorithms and introduces SLIC to evaluate boundary adherence, speed, memory usage, and segmentation impact. The authors empirically compare the algorithms on boundary adherence, speed, memory, and segmentation performance, and present SLIC as a k‑means‑based clustering method for efficient superpixel generation. SLIC matches or surpasses prior methods in boundary adherence, while being faster, more memory efficient, improving segmentation performance, and easily extendable to supervoxel generation.

Abstract

Computer vision applications have come to rely increasingly on superpixels in recent years, but it is not always clear what constitutes a good superpixel algorithm. In an effort to understand the benefits and drawbacks of existing methods, we empirically compare five state-of-the-art superpixel algorithms for their ability to adhere to image boundaries, speed, memory efficiency, and their impact on segmentation performance. We then introduce a new superpixel algorithm, simple linear iterative clustering (SLIC), which adapts a k-means clustering approach to efficiently generate superpixels. Despite its simplicity, SLIC adheres to boundaries as well as or better than previous methods. At the same time, it is faster and more memory efficient, improves segmentation performance, and is straightforward to extend to supervoxel generation.