Scene motion, multiple reflections, and sensor noise introduce artifacts in the depth reconstruction performed by time-of-flight (ToF) cameras.

We propose a two-stage, deep-learning approach to address all of these sources of artifacts simultaneously.

We also introduce FLAT, a synthetic dataset of 2000 ToF measurements that capture all of these nonidealities, and allows to simulate different camera hardware. Using the Kinect 2 camera as a baseline, we show improved reconstruction errors over state-of-the-art methods, on both simulated and real data.

We describe a new technique for obtaining a spatially varying BRDF (svBRDF) of a flat object using printed fiducial markers and a cell phone capable of continuous flash video. Our homography-based video frame alignment method does not require the fiducial markers to be visible in every frame, thereby enabling us to capture larger areas at a closer distance and higher resolution than in previous work.

Given two consecutive frames, video interpolation aims at generating intermediate frame(s) to form both spatially and temporally coherent video sequences. While most existing methods focus on single-frame interpolation, we propose an end-to-end convolutional neural network for variable-length multi-frame video interpolation, where the motion interpretation and occlusion reasoning are jointly modeled. We start by computing bi-directional optical flow between the input images using a U-Net architecture.

Point cloud registration sits at the core of many important and challenging 3D perception problems including autonomous navigation, SLAM, object/scene recognition, and augmented reality. In this paper, we present a new registration algorithm that is able to achieve state-of-the-art speed and accuracy through its use of a hierarchical Gaussian Mixture Model (GMM) representation. Our method constructs a top-down multi-scale representation of point cloud data by recursively running many small-scale data likelihood segmentations in parallel on a GPU.