Introduction

As a pre-processing step for many problems in the field of computer vision, superpixel algorithms aim to over-segment the image by grouping homogenous pixels. In this paper, we propose a novel superpixel segmentation algorithm, namely Superpixel using the Shortest Gradient Distance (SSGD for short) in a k-means clustering framework. Starting from initializing the superpixel seeds, a bilateral filter is applied to the texture-rich regions centered at each initial seed. Then, a new distance function taking the shortest gradient distance into account is computed to enforce adherence to boundaries. Unlike using the simple Euclidean distance, the proposed combined distance function increases the accuracy of associating a pixel to a cluster. The experimental results demonstrate that our algorithm outperforms the state-of-the-art methods.

Image reconstruction with 300 superpixels using mean color of clusters. (a) original image, (b) SLIC[1], (c) SCALP[2], (d) SSGD without bilateral filtering operation, (e) SSGD.

Source Code

https://github.com/noahzn/SSGD

Experiment Results

We compare SSGD with several state-of-the-art methods, including SLIC [1], SCALP [2], TP [3], NC [4], on BSDS500 [5].

We use three standard metrics to evaluate the performance of different algorithms, which are Boundary Recall (BR), Corrected Undersegmentation Error (CUE) and Achievable Segmentation Accuracy (ASA). For BR and ASA the higher the better, while for CUE the lower the better.

Reference 

[1] R. Achanta, A. Shaji, K. Smith, et al., “Slic superpixels compared to state-of-the-art superpixel methods,” IEEE Trans. PAMI, vol. 34, no. 11, pp. 2274–82, 2012.

[2] R. Giraud, V. Ta, and N. Papadakis, “Scalp: Superpixels with contour adherence using linear path,” in ICPR, 2016.

[3] A. Levinshtein, A. Stere, K. N. Kutulakos, et al., “Turbopixels: Fast superpixels using geometric flows,” IEEE Trans. PAMI, vol. 31, no. 12, pp. 2290–7, 2009.

[4] J. Shi and J. Malik, “Normalized cuts and image segmentation,” IEEE Trans. PAMI, vol. 22, no. 8, pp. 888–905, 2000.

[5] P. Arbelaez, M. Maire, C. Fowlkes C, et al., “Contour detection and hierarchical image segmentation,” IEEE Trans. PAMI, vol. 33, no. 5, pp. 898–916, 2011.