We present a technique for computing a dense pixel correspondence between two images of a scene containing multiple large, rigid motions. We model each motion with either a homography (for planar objects) or a fundamental matrix. The various motions in the scene are first extracted by clustering an initial sparse set of correspondences between feature points; we then perform a multi-label graph cut optimization which assigns each pixel to an independent motion and computes its disparity with respect to that motion. We demonstrate our technique on several example scenes and compare our results with previous approaches.
Dense pixel correspondence between two images of a scene with moving rigid objects. Images (a) and (b) are the input, and (c) is the flow field from (a) to (b) with color labelings. Red indicates occluded areas, and blue and green indicate two motions modeled by fundamental matrices for the bird and the rest of the background(note that the flow vectors are uniformly scaled to a shorter length for visual clarity). Image (d) is the reconstruction of (a) based on the flow; red areas are occluded.