Pulling Things out of Perspective

By Lubor Ladicky, Jianbo Shi, Marc Pollefeys. 2014

492 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition, Media Technology, 3D Vision & Pose

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

In Media Technology

In 3D Vision & Pose

Citations in the corpus

The Cityscapes Dataset for Semantic Urban Scene Understanding (2016, background): “As no official pixel-wise annotations exist for KITTI, several independent groups have annotated approximately 700 frames [22, 29, 32, 33, 58, 64, 77, 80].”
Conditional Random Fields as Recurrent Neural Networks (2015, background): “[15] recently showed that it is possible to employ a discriminative supervised approach to simultaneously predict pixelwise labels with semantic image segmentation, depth estimation, and surface normal [16].”
Predicting Depth, Surface Normals and Semantic Labels with a Common Multi-scale Convolutional Architecture (2014, methodology): “We combine both depths and normals networks together to share computation, creating a network using a single scale 1 stack, but separate scale 2 and 3 stacks.”
Generative Image Modeling Using Style and Structure Adversarial Networks (2016, methodology): “More specifically, we quantize the surface normals to 40 classes with k-means clustering as in [39,54] and the loss is defined as”
Learning a Multi-View Stereo Machine (2017, background): “Approaches which predict shape from a single image have been proposed in form of direct depth map regression [46, 31, 10], generating multiple depth maps from novel viewpoints [51], producing voxel occupancies [5, 16], geometry images [49] and point clouds [11].”
Hierarchical Surface Prediction for 3D Object Reconstruction (2017, background): “5D setting where the goal is, given a color image to predict a depth map [11], [21], [30].”
Lifting Object Detection Datasets into 3D (2015, positive): “To account for differences between the 3D exemplars and the object depicted, Su and Guibas [26] choose a few exemplars that appear to be similar to the depicted shape and combine them to produce a single depth map.”
Depth Map Prediction from a Single Image using a Multi-Scale Deep Network (2014, methodology): “[12] show how to integrate semantic object labels with monocular depth features to improve performance; however, they rely on handcrafted features and use superpixels to segment the image.”
What Uncertainties Do We Need in Bayesian Deep Learning for Computer Vision? (2017, neutral): “certainty 0.162 3.87 0.064 + Aleatoric & Epistemic 0.149 4.08 0.063 (a) Make3D depth dataset [25]. NYU v2 Depth rel rms log 10 1 2 3 Karsch et al. [33] 0.374 1.12 0.134 - - - Ladicky et al. [36] - - - 54.2% 82.9% 91.4% Liu et al. [34] 0.335 1.06 0.127 - - - Li et al. [35] 0.232 0.821 0.094 62”
Deep convolutional neural fields for depth estimation from a single image (2014, methodology): “Most notably, our results are significantly better than that of [8], which jointly exploits depth estimation and semantic labelling.”
Full-Resolution Residual Networks for Semantic Segmentation in Street Scenes (2016, background): “Semantic image segmentation [49, 38, 21, 53, 33], the task of assigning a set of predefined class labels to image pixels, is an important tool for modeling the complex relationships of the semantic entities usually found in street scenes, such as cars, pedestrians, road, or sidewalks.”
Deeper Depth Prediction with Fully Convolutional Residual Networks (2016, methodology): “[15] instead jointly predict labels and depths in a classification approach.”
Unsupervised CNN for Single View Depth Estimation: Geometry to the Rescue (2016, background): “n labels for Imagenet requires signicant human eort and is prone to error. Supervised training for single view depth estimation for outdoor scenes requires expensive hardware and careful acquisition[29,8,24,21]. 1 Find the model and other imformation on the project github page: https://github. com/R”
Refine and Distill: Exploiting Cycle-Inconsistency and Knowledge Distillation for Unsupervised Monocular Depth Estimation (2019, background): “In the last few years, deep learning-based approaches for depth estimation [5, 21, 25, 38, 13, 43, 28, 32] have attracted a growing interest, motivated, on the one hand, by their ability to predict very accurate depth maps and, on the other hand, by the importance of recovering depth information in ”
Revisiting Single Image Depth Estimation: Toward Higher Resolution Maps With Accurate Object Boundaries (2018, background): “ages; those with ∗ use relative depth between pairs or partially known depths [4, 28], and those with ∗∗ employ joint task learning [20, 37, 23, 8, 6, 39].”
Enforcing Geometric Constraints of Virtual Normal for Depth Prediction (2019, unknown): “[26] - - - 0.”
Joint Task-Recursive Learning for Semantic Segmentation and Depth Estimation (2018, background): “It involves many cross tasks, such as detection and classification [16] [17], depth estimation and image decomposition [18], image segmentation and classification [19], and also depth estimation and semantic segmentation [20] [21] [22], etc.”
Multi-scale Continuous CRFs as Sequential Deep Networks for Monocular Depth Estimation (2017, background): “Ladicky [17] demonstrated the beneﬁt of combining semantic object labels with depth features.”