We propose an end-to-end deep learning architecture that produces a 3D shape in triangular mesh from a single color image. Limited by the nature of deep neural network, previous methods usually represent a 3D shape in volume or point cloud, and it is non-trivial to convert them to the more ready-to-use mesh model. Unlike the existing methods, our network represents 3D mesh in a graph-based convolutional neural network and produces correct geometry by progressively deforming an ellipsoid, leveraging perceptual features extracted from the input image. We adopt a coarse-to-fine strategy to make the whole deformation procedure stable, and define various of mesh related losses to capture properties of different levels to guarantee visually appealing and physically accurate 3D geometry. Extensive experiments show that our method not only qualitatively produces mesh model with better details, but also achieves higher 3D shape estimation accuracy compared to the state-of-the-art.

Access to large, diverse RGB-D datasets is critical for training RGB-D scene understanding algorithms. However, existing datasets still cover only a limited number of views or a restricted scale of spaces. In this paper, we introduce Matterport3D, a large-scale RGB-D dataset containing 10,800 panoramic views from 194,400 RGB-D images of 90 building-scale scenes. Annotations are provided with surface reconstructions, camera poses, and 2D and 3D semantic segmentations. The precise global alignment and comprehensive, diverse panoramic set of views over entire buildings enable a variety of supervised and self-supervised computer vision tasks, including keypoint matching, view overlap prediction, normal prediction from color, semantic segmentation, and region classification.

No abstract

The goal of our work is to complete the depth channel of an RGB-D image. Commodity-grade depth cameras often fail to sense depth for shiny, bright, transparent, and distant surfaces. To address this problem, we train a deep network that takes an RGB image as input and predicts dense surface normals and occlusion boundaries. Those predictions are then combined with raw depth observations provided by the RGB-D camera to solve for depths for all pixels, including those missing in the original observation. This method was chosen over others (e.g., inpainting depths directly) as the result of extensive experiments with a new depth completion benchmark dataset, where holes are filled in training data through the rendering of surface reconstructions created from multiview RGB-D scans. Experiments with different network inputs, depth representations, loss functions, optimization methods, inpainting methods, and deep depth estimation networks show that our proposed approach provides better depth completions than these alternatives. Depth representation:The obvious approach to address our problem is to use the new dataset as supervision to train a fully convolutional network to regress depth directly from RGB-D. However, that approach does not work very well, especially for large holes like the one shown in the bottom row of Figure 1. Estimating absolute depths from a monocular color image is difficult even for people [53]. Rather, we train the network to predict only local differential properties of depth (surface normals and occlusion boundaries), which are much easier to estimate [35]. We then solve for the absolute depths with a global optimization. Deep network design:There is no previous work on studying how best to design and train an end-to-end deep network for completing depth images from RGB-D inputs. At first glance, it seems straight-forward to extend previous net-1 arXiv:1803.09326v2 [cs.CV]

In this paper, we propose a deep learning architecture that produces accurate dense depth for the outdoor scene from a single color image and a sparse depth. Inspired by the indoor depth completion, our network estimates surface normals as the intermediate representation to produce dense depth, and can be trained end-to-end. With a modified encoder-decoder structure, our network effectively fuses the dense color image and the sparse LiDAR depth. To address outdoor specific challenges, our network predicts a confidence mask to handle mixed LiDAR signals near foreground boundaries due to occlusion, and combines estimates from the color image and surface normals with learned attention maps to improve the depth accuracy especially for distant areas. Extensive experiments demonstrate that our model improves upon the state-of-the-art performance on KITTI depth completion benchmark. Ablation study shows the positive impact of each model components to the final performance, and comprehensive analysis shows that our model generalizes well to the input with higher sparsity or from indoor scenes. * indicates equal contributions.† indicates corresponding author. Color Image Sparse Data from LiDARDeepLiDAR: Our Dense Prediction (colored with input color image)DeepLiDAR: Our Dense Prediction (colored with surface normal)

Indoor scene understanding is central to applications such as robot navigation and human companion assistance. Over the last years, data-driven deep neural networks have outperformed many traditional approaches thanks to their representation learning capabilities. One of the bottlenecks in training for better representations is the amount of available per-pixel ground truth data that is required for core scene understanding tasks such as semantic segmentation, normal prediction, and object boundary detection. To address this problem, a number of works proposed using synthetic data. However, a systematic study of how such synthetic data is generated is missing. In this work, we introduce a large-scale synthetic dataset with 500K physicallybased rendered images from 45K realistic 3D indoor scenes. We study the effects of rendering methods and scene lighting on training for three computer vision tasks: surface normal prediction, semantic segmentation, and object boundary detection. This study provides insights into the best practices for training with synthetic data (more realistic rendering is worth it) and shows that pretraining with our new synthetic dataset can improve results beyond the current state of the art on all three tasks.

Abstract. The field-of-view of standard cameras is very small, which is one of the main reasons that contextual information is not as useful as it should be for object detection. To overcome this limitation, we advocate the use of 360• full-view panoramas in scene understanding, and propose a whole-room context model in 3D. For an input panorama, our method outputs 3D bounding boxes of the room and all major objects inside, together with their semantic categories. Our method generates 3D hypotheses based on contextual constraints and ranks the hypotheses holistically, combining both bottom-up and top-down context information. To train our model, we construct an annotated panorama dataset and reconstruct the 3D model from single-view using manual annotation. Experiments show that solely based on 3D context without any image-based object detector, we can achieve a comparable performance with the state-of-the-art object detector. This demonstrates that when the FOV is large, context is as powerful as object appearance. All data and source code are available online.

No abstract