Superpixel-Based Background Recovery from Multiple Images

TL;DR

This paper introduces a superpixel-based method for recovering backgrounds from multiple images, involving segmentation, iterative updating, and artifact removal, demonstrated effectively on outdoor datasets.

Contribution

The paper presents a novel superpixel segmentation approach combined with iterative model updating for background recovery from multiple images.

Findings

Achieves promising background recovery results on outdoor datasets.

Effectively removes ghosting artifacts with k-nearest neighbor method.

Demonstrates robustness in handling dynamic scenes with minimal background change.

Abstract

In this paper, we propose an intuitive method to recover background from multiple images. The implementation consists of three stages: model initialization, model update, and background output. We consider the pixels whose values change little in all input images as background seeds. Images are then segmented into superpixels with simple linear iterative clustering. When the number of pixels labelled as background in a superpixel is bigger than a predefined threshold, we label the superpixel as background to initialize the background candidate masks. Background candidate images are obtained from input raw images with the masks. Combining all candidate images, a background image is produced. The background candidate masks, candidate images, and the background image are then updated alternately until convergence. Finally, ghosting artifacts is removed with the k-nearest neighbour method. An experiment on an outdoor dataset demonstrates that the proposed algorithm can achieve promising results.