Depth Map Super-Resolution by Deep Multi-Scale Guidance

By Tak-Wai Hui, Chen Change Loy, Xiaoou Tang. 2016

373 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition, Media Technology

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

Citations in the corpus

Multi-View Silhouette and Depth Decomposition for High Resolution 3D Object Representation (2018, background): “The shape of a 3D object can be completely captured by a set of 2D images from multiple viewpoints (see [21, 4] for an analysis of selecting the location and number of viewpoints).”
Sparsity Invariant CNNs (2017, background): “Several approaches also exploit end-to-end models for guided depth upsampling of regular data [28,61].”
NTIRE 2017 Challenge on Single Image Super-Resolution: Methods and Results (2017, methodology): “The upsampling operation is performed by the deconvolutional layers [10, 15] and a multi-kernel upsampling is used (see Fig.”
Accelerating the Super-Resolution Convolutional Neural Network (2016, background): “Motivated by SRCNN, some problems such as face hallucination [16] and depth map super-resolution [17] have achieved state-of-the-art results.”
ARKitScenes: A Diverse Real-World Dataset For 3D Indoor Scene Understanding Using Mobile RGB-D Data (2021, methodology): “For these experiments we reproduced three classical approaches for depth upsampling - naïve Bilinear interpolation, Joint Bilateral Upsampling (JBU) [38] and Fast Guided Global Interpolation (FGI) [39], as well as the two mentioned modern DNN-based solutions - MSG [37] and MSPF [45].”
Deep Learning for Image Super-Resolution: A Survey (2019, methodology): “[166] utilize two CNNs to simultaneously upsample LR depth maps and downsample HR RGB images, then use RGB features as the guidance for upsampling depth maps with the same resolution.”
Learning Depth with Convolutional Spatial Propagation Network (2018, background): “Recently, deep learning based enhancement yields impressive results on super resolution of both images [61], [62] and depths [63], [64], [65], [66].”
Super-Resolution for Remote Sensing Images via Local–Global Combined Network (2017, background): “When a network goes deep, learning residuals can make the network converge faster and obtain a better minimum and performance [15][16][17].”
Depth Estimation via Affinity Learned with Convolutional Spatial Propagation Network (2018, background): “Recently, deep learning based enhancement yields impressive results on super resolution of both images [33,34] and depths [35,36,37,38].”
Recovering Realistic Texture in Image Super-Resolution by Deep Spatial Feature Transform (2018, background): “Multiscale guidance structure has also been proposed for depth map super-resolution [18].”
In Defense of Classical Image Processing: Fast Depth Completion on the CPU (2018, background): “Recently, deep learning approaches have emerged to tackle the guided depth completion problem [15], [16].”