Abstract

We present a new parsing framework for the line-based geo- metric analysis of a single image coming from a man-made environment. This parsing framework models the scene as a composition of geomet- ric primitives spanning different layers from low level (edges) through mid-level (lines and vanishing points) to high level (the zenith and the horizon). The inference in such a model thus jointly and simultaneously estimates a) the grouping of edges into the straight lines, b) the grouping of lines into parallel families, and c) the positioning of the horizon and the zenith in the image. Such a unified treatment means that the un- certainty information propagates between the layers of the model. This is in contrast to most previous approaches to the same problem, which either ignore the middle levels (lines) all together, or use the bottom-up step-by-step pipeline. For the evaluation, we consider a publicly available York Urban dataset of “Manhattan” scenes, and also introduce a new, harder dataset of 103 urban outdoor images containing many non-Manhattan scenes. The comparative evaluation for the horizon estimation task demonstrate higher accuracy and robustness attained by our method when compared to the current state-of-the-art approaches.