GAL: A global-attributes assisted labeling system for outdoor scenes

- A three-stage geometric labeling system which integrates local information and global information is proposed.
- Seven global attributes for geometric labeling are proposed.
- The proposed system is able to handle a large variety of scenes, both urban scene and natural scene.
- A graphical model is proposed to integrate local and global information and to optimize geometric labeling.
- Our system achieves state-of-the art performance against a popular benchmark dataset.

An approach that extracts global attributes from outdoor images to facilitate geometric layout labeling is investigated in this work. The proposed Global-attributes Assisted Labeling (GAL) system exploits both local features and global attributes. First, by following a classical method, we use local features to provide initial labels for all super-pixels. Then, we develop a set of techniques to extract global attributes from 2D outdoor images. They include sky lines, ground lines, vanishing lines, etc. Finally, we propose the GAL system that integrates global attributes in the conditional random field (CRF) framework to improve initial labels so as to offer a more robust labeling result. The performance of the proposed GAL system is demonstrated and benchmarked with several state-of-the-art algorithms against a popular outdoor scene layout dataset.