Abstract:Depth estimation from monocular images is an ill-posed and inherently ambiguous problem. Recently, deep learning technique has been applied for monocular depth estimation seeking data-driven solutions. However, most existing methods focus on pursuing the minimization of average depth regression error at pixel level and neglect to encode the global layout of scene, resulting in layout-inconsistent depth map. This paper proposes a novel Layout-Aware Convolutional Neural Network (LA-Net) for accurate monocular depth estimation by simultaneously perceiving scene layout and local depth details. Specifically, a Spatial Layout Network (SL-Net) is proposed to learn a layout map representing the depth ordering between local patches. A Layout-Aware Depth Estimation Network (LDE-Net) is proposed to estimate pixel-level depth details using multi-scale layout maps as structural guidance, leading to layout-consistent depth map. A dense network module is used as the base network to learn effective visual details resorting to dense feed-forward connections. Moreover, we formulate an order-sensitive softmax loss to well constrain the ill-posed depth inferring problem. Extensive experiments on both indoor scene (NYUD-v2) and outdoor scene (Make3D) datasets have demonstrated that the proposed LA-Net outperforms the state-of-the-art methods and leads to faithful 3D projections.

Room Layout Estimation by Learning Depth Maps of Planes from 2D Layout Labels

GeoLayout: Geometry Driven Room Layout Estimation Based on Depth Maps of Planes

Edge-Semantic Learning Strategy for Layout Estimation in Indoor Environment

3D Room Layout Estimation from a Single RGB Image.

Structural Deep Metric Learning for Room Layout Estimation

RoomStructNet: Learning to Rank Non-Cuboidal Room Layouts From Single View

Self-supervised 360$^{\circ}$ Room Layout Estimation

Web3d Learning Platform of Furniture Layout Based on Case-Based Reasoning and Distance Field.

Learning to Reconstruct 3D Non-Cuboid Room Layout from a Single RGB Image

Bottom-up Estimation of Geometric Layout for Indoor Images

Spatial Geometric Reasoning for Room Layout Estimation via Deep Reinforcement Learning

Estimating spatial layout of rooms from RGB-D videos.

Self-training Room Layout Estimation via Geometry-aware Ray-casting

LED2-Net: Monocular 360 Layout Estimation via Differentiable Depth Rendering

Low-Drift RGB-D SLAM with Room Reconstruction Using Scene Understanding

Estimating Generic 3D Room Structures from 2D Annotations

3D Room Layout Estimation from a Cubemap of Panorama Image Via Deep Manhattan Hough Transform

From Edge to Keypoint: An End-to-End Framework For Indoor Layout Estimation

LayoutNet: Reconstructing the 3D Room Layout from a Single RGB Image

Indoor Layout Estimation by 2D LiDAR and Camera Fusion

LA-Net: Layout-Aware Dense Network for Monocular Depth Estimation.