Deep Motion Boundary Detection
This addresses a crucial problem in computer vision for applications like video analysis, though it is incremental as it builds on existing optical flow methods.
The authors tackled motion boundary detection by proposing MoBoNet, the first end-to-end deep learning approach, which uses a refinement network with images, optical flows, and warping errors as inputs to produce high-resolution motion boundaries and refine optical flows, achieving results superior to state-of-the-art methods.
Motion boundary detection is a crucial yet challenging problem. Prior methods focus on analyzing the gradients and distributions of optical flow fields, or use hand-crafted features for motion boundary learning. In this paper, we propose the first dedicated end-to-end deep learning approach for motion boundary detection, which we term as MoBoNet. We introduce a refinement network structure which takes source input images, initial forward and backward optical flows as well as corresponding warping errors as inputs and produces high-resolution motion boundaries. Furthermore, we show that the obtained motion boundaries, through a fusion sub-network we design, can in turn guide the optical flows for removing the artifacts. The proposed MoBoNet is generic and works with any optical flows. Our motion boundary detection and the refined optical flow estimation achieve results superior to the state of the art.