Abstract
We present a probabilistic, online, depth map fusion frame- work, whose generative model for the sensor measurement process accu- rately incorporates both long-range visibility constraints and a spatially varying, probabilistic outlier model. In addition, we propose an inference algorithm that updates the state variables of this model in linear time each frame. Our detailed evaluation compares our approach against sev- eral others, demonstrating and explaining the improvements that this model offers, as well as highlighting a problem with all current methods: systemic bias.