Make3D: Learning 3D Scene Structure from a Single Still Image

Saxena, Ashutosh; Sun, Min; Ng, Andrew Y.

doi:10.1109/tpami.2008.132

Cited by 1,572 publications

(1,184 citation statements)

References 38 publications

Supporting

Mentioning

1,174

Contrasting

Unclassified

Order By: Relevance

“…laser scanning, are not convenient for collecting large amounts of high quality data in varied environments. Some range datasets do exist online, such as the Brown Range Image Database [24], the USF Range Database [33] and Make3D Range Image Database [34]. The USF dataset is a collection of 400 range images of simple polyhedral objects with a very limited resolution of only 128 × 128 pixels and with heavily quantized depth.…”

Section: Training Datamentioning

confidence: 99%

Patch Based Synthesis for Single Depth Image Super-Resolution

Aodha

Campbell

Nair

et al. 2012

Computer Vision – ECCV 2012

163

196

View full text Add to dashboard Cite

Abstract. We present an algorithm to synthetically increase the resolution of a solitary depth image using only a generic database of local patches. Modern range sensors measure depths with non-Gaussian noise and at lower starting resolutions than typical visible-light cameras. While patch based approaches for upsampling intensity images continue to improve, this is the first exploration of patching for depth images. We match against the height field of each low resolution input depth patch, and search our database for a list of appropriate high resolution candidate patches. Selecting the right candidate at each location in the depth image is then posed as a Markov random field labeling problem. Our experiments also show how important further depth-specific processing, such as noise removal and correct patch normalization, dramatically improves our results. Perhaps surprisingly, even better results are achieved on a variety of real test scenes by providing our algorithm with only synthetic training depth data.

show abstract

Section: Training Datamentioning

confidence: 99%

Patch Based Synthesis for Single Depth Image Super-Resolution

Aodha

Campbell

Nair

et al. 2012

Computer Vision – ECCV 2012

163

196

View full text Add to dashboard Cite

show abstract

“…Notable examples are [14,30,22,33,17]. These focus on single images and, unlike our work, they do not attempt to enforce consistency across views.…”

Section: Related Workmentioning

confidence: 99%

Semantic Structure from Motion: A Novel Framework for Joint Object Recognition and 3D Reconstruction

Bao

Savarese

2012

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Abstract. Conventional rigid structure from motion (SFM) addresses the problem of recovering the camera parameters (motion) and the 3D locations (structure) of scene points, given observed 2D image feature points. In this chapter, we propose a new formulation called Semantic Structure From Motion (SSFM). In addition to the geometrical constraints provided by SFM, SSFM takes advantage of both semantic and geometrical properties associated with objects in a scene. These properties allow to jointly estimate the structure of the scene, the camera parameters as well as the 3D locations, poses, and categories of objects in a scene. We cast this problem as a max-likelihood problem where geometry (cameras, points, objects) and semantic information (object classes) are simultaneously estimated. The key intuition is that, in addition to image features, the measurements of objects across views provide additional geometrical constraints that relate cameras and scene parameters. These constraints make the geometry estimation process more robust and, in turn, make object detection more accurate. Our framework has the unique ability to: i) estimate camera poses only from object detections, ii) enhance camera pose estimation, compared to feature-pointbased SFM algorithms, iii) improve object detections given multiple uncalibrated images, compared to independently detecting objects in single images. Extensive quantitative results on three datasets -LiDAR cars, street-view pedestrians, and Kinect office desktop -verify our theoretical claims.

show abstract

“…[31,9] predict the depth from a single image by learning models from training data. Single-view reconstruction has been proposed for robot navigation and planning [26,32], but its accuracy is usually lower than multiview techniques and might fail catastrophically if the underlying assumptions are not met or the current image is far from the training set.…”

Section: Introductionmentioning

confidence: 99%

Automatic implementation of the numerical Taylor series method: A Mathematica and Sage approach

Abad

Barrio

Marco-Buzunáriz

et al. 2015

Applied Mathematics and Computation

View full text Add to dashboard Cite

This paper presents a dense monocular mapping algorithm that improves the accuracy of the state-of-the-art variational and multiview stereo methods by incorporating scene priors into its formulation. Most of the improvement of our proposal is in lowtextured image regions and for low-parallax camera motions; two typical failure cases of multiview mapping.The specific priors we model are the planarity of homogeneous color regions, the repeating geometric primitives of the scene -that can be learned from data-and the Manhattan structure of indoor rooms. We evaluate the performance of our method in our own sequences and in the publicly available NYU dataset, emphasizing its strengths and weaknesses in different cases.

show abstract

Make3D: Learning 3D Scene Structure from a Single Still Image

Cited by 1,572 publications

References 38 publications

Patch Based Synthesis for Single Depth Image Super-Resolution

Patch Based Synthesis for Single Depth Image Super-Resolution

Semantic Structure from Motion: A Novel Framework for Joint Object Recognition and 3D Reconstruction

Automatic implementation of the numerical Taylor series method: A Mathematica and Sage approach

Contact Info

Product

Resources

About