We propose a simple but powerful multi-view semantic segmentation framework for images captured by a camera mounted on a car driving along streets. In our approach, a pair-wise Markov Random Field (MRF) is laid out across multiple views. Both 2D and 3D features are extracted at a super-pixel level to train classifiers for the unary data terms of MRF. For smoothness terms, our approach makes use of color differences in the same image to identify accurate segmentation boundaries, and dense pixel-to-pixel correspondences to enforce consistency across different views. To speed up training and to improve the recognition quality, our approach adaptively selects the most similar training data for each scene from the label pool. Furthermore, we also propose a powerful approach within the same framework to enable large-scale labeling in both the 3D space and 2D images. We demonstrate our approach on more than 10,000 images from Google Maps Street View.
Link to High Resolution Poster (126M)
This work was supported by Hong Kong RGC Grants 618908, 619107, 619006, and RGC/NSFC NHKUST602/05. We thank Qiang Bi for labeling some data and the anonymous reviewers and the area chair for constructive comments that helped to improve this work. The data set was kindly provided by Google.