2,173 Hits in 3.2 sec

Depth Map Super-Resolution Using Guided Deformable Convolution

Joon-Yeon Kim, Seowon Ji, Seung-Jin Baek, Seung-Won Jung, Sung-Jea Ko
2021 IEEE Access  
KIM et al.: Depth Map Super-Resolution Based on Guided Deformable Convolution  ...  Observing that recent color-guided depth map SR networks produce unwanted texture copying artifacts because of the excessive use of guidance features, we proposed guided deformable convolution, which uses  ... 
doi:10.1109/access.2021.3076853 fatcat:qd6k7h2cujghnoqxavtvb7bj24

Content-aware Directed Propagation Network with Pixel Adaptive Kernel Attention [article]

Min-Cheol Sagong, Yoon-Jae Yeo, Seung-Won Jung, Sung-Jea Ko
2021 arXiv   pre-print
We also show the generalizability of the proposed method by applying it to multi-modal tasks especially color-guided depth map super-resolution.  ...  The proposed method infers pixel-adaptive attention maps along the channel and spatial directions separately to address the decomposed model with fewer parameters.  ...  depth map super-resolution.  ... 
arXiv:2107.13144v1 fatcat:zgkf4vz7iffmjpzsoq4drigty4

Single Pair Cross-Modality Super Resolution [article]

Guy Shacht, Sharon Fogel, Dov Danon, Daniel Cohen-Or, Ilya Leizerson
2021 arXiv   pre-print
To this end, Cross-Modality Super-Resolution methods were introduced, where an RGB image of a high-resolution assists in increasing the resolution of the low-resolution modality.  ...  Non-visual imaging sensors are widely used in the industry for different purposes. Those sensors are more expensive than visual (RGB) sensors, and usually produce images with lower resolution.  ...  In Guided Super-Resolution as Pixel-to-Pixel Transformation ( [13] ), the problem of guided depth-maps SR was posed as a pixel-to-pixel translation of the HR guiding modality to a newly predicted HR depth-map  ... 
arXiv:2004.09965v4 fatcat:56d7xmsidjfffpci62dd4lgmuy

FreqNet: A Frequency-domain Image Super-Resolution Network with Dicrete Cosine Transform [article]

Runyuan Cai, Yue Ding, Hongtao Lu
2021 arXiv   pre-print
Single image super-resolution(SISR) is an ill-posed problem that aims to obtain high-resolution (HR) output from low-resolution (LR) input, during which extra high-frequency information is supposed to  ...  We further observe that our model can be merged with other spatial super-resolution models to enhance the quality of their original SR output.  ...  Effects of Deformable(DRG) and Depth-wise Residual Group(DWRG).  ... 
arXiv:2111.10800v1 fatcat:e3lgbaaehzetlcz5cavjhrlo3a

Learning Graph Regularisation for Guided Super-Resolution [article]

Riccardo de Lutio and Alexander Becker and Stefano D'Aronco and Stefania Russo and Jan D. Wegner and Konrad Schindler
2022 arXiv   pre-print
We introduce a novel formulation for guided super-resolution. Its core is a differentiable optimisation layer that operates on a learned affinity graph.  ...  With the decision to employ the source as a constraint rather than only as an input to the prediction, our method differs from state-of-the-art deep architectures for guided super-resolution, which produce  ...  depth map.  ... 
arXiv:2203.14297v1 fatcat:jrvlo7vmyncfbdbntkxdo3j7v4

RGB Guided Depth Map Super-Resolution with Coupled U-Net

Yingjie Cui, Qingmin Liao, Wenming Yang, Jing-Hao Xue
2021 2021 IEEE International Conference on Multimedia and Expo (ICME)  
The depth maps captured by RGB-D cameras usually are of low resolution, entailing recent efforts to develop depth super-resolution (DSR) methods.  ...  Secondly, high-resolution (HR) RGB features and low-resolution (LR) depth features are often fused in shallow layers only. Thirdly, only the last layer of features is used for reconstruction.  ...  CONCLUSIONS In this paper, we propose Coupled U-Net (CU-Net) for color guided depth map super-resolution.  ... 
doi:10.1109/icme51207.2021.9428096 fatcat:5jubanzrxzbwlkm55mftf45jki

Guided Cascaded Super-Resolution Network for Face Image

Lin Cao, Jiape Liu, Kangning Du, Yanan Guo, Tao Wang
2020 IEEE Access  
It stacks three convolutional layers to learn the non-linear mapping relationship between HR and LR image pairs to achieve image super-resolution reconstruction.  ...  Finally, a single convolutional layer recovers the feature map to the high-resolution image.  ... 
doi:10.1109/access.2020.3025972 fatcat:nghrwibck5bsngps5jfr7lgvd4

Deformable Non-local Network For Video Super-Resolution [article]

Hua Wang, Dewei Su, Longcun Jin, Chuangchuang Liu
2019 arXiv   pre-print
The video super-resolution (VSR) task aims to restore a high-resolution video frame by using its corresponding low-resolution frame and multiple neighboring frames.  ...  Specifically, we apply the improved deformable convolution in our alignment module to achieve adaptive frame alignment at the feature level.  ...  VDSR [10] further improved SRCNN by stacking more convolution layers and using residual learning to increase network depth.  ... 
arXiv:1909.10692v1 fatcat:kvmdpw3gpjhyvitfj7cjhc7e4e

Depth Map Super-Resolution via Cascaded Transformers Guidance

Ido Ariav, Israel Cohen
2022 Frontiers in Signal Processing  
Several methods have recently been proposed for guided super-resolution of depth maps using convolutional neural networks to overcome this limitation.  ...  In a guided super-resolution scheme, high-resolution depth maps are inferred from low-resolution ones with the additional guidance of a corresponding high-resolution intensity image.  ...  Intensity Guided Depth Map Super Resolution Unlike an HR depth map, an HR intensity image can usually be easily acquired by color cameras.  ... 
doi:10.3389/frsip.2022.847890 fatcat:bzrqbsx72fgcvoir267666onpe

DeepPyram: Enabling Pyramid View and Deformable Pyramid Reception for Semantic Segmentation in Cataract Surgery Videos [article]

Negin Ghamsarian, Mario Taschwer, klaus Schoeffmann
2021 arXiv   pre-print
feature map; (ii) Deformable Pyramid Reception, which enables a wide deformable receptive field that can adapt to geometric transformations in the object of interest; and (iii) Pyramid Loss that adaptively  ...  supervises multi-scale semantic feature maps.  ...  In the second case, the loss is computed by performing super-resolution through interpolating the output feature map (to obtain a feature map with the same resolution as the ground truth) and comparing  ... 
arXiv:2109.05352v1 fatcat:n5u6crsgszblfgxnuxobzzgy6i

Cross-MPI: Cross-scale Stereo for Image Super-Resolution using Multiplane Images [article]

Yuemei Zhou, Gaochang Wu, Ying Fu, Kun Li, Yebin Liu
2021 arXiv   pre-print
Specifically, we propose Cross-MPI, an end-to-end RefSR network composed of a novel plane-aware attention-based MPI mechanism, a multiscale guided upsampling module as well as a super-resolution (SR) synthesis  ...  However, existing RefSR approaches fail to accomplish high-fidelity super-resolution under a large resolution gap, e.g., 8x upscaling, due to the lower consideration of the underlying scene structure.  ...  To match the frequency band of I LR and obtain alpha maps with target resolution and fine details, we first calculate rough alpha maps at low resolution using planeaware attention as depicted in Sect.  ... 
arXiv:2011.14631v2 fatcat:cmkve6widfcv3ohx3zyswqye34

Video Super Resolution Based on Deep Learning: A Comprehensive Survey [article]

Hongying Liu, Zhubo Ruan, Peng Zhao, Chao Dong, Fanhua Shang, Yuanyuan Liu, Linlin Yang, Radu Timofte
2022 arXiv   pre-print
It is well known that the leverage of information within video frames is important for video super-resolution.  ...  In this survey, we comprehensively investigate 33 state-of-the-art video super-resolution (VSR) methods based on deep learning.  ...  The offsets are applied to the conventional convolution kernel to yield a deformable convolution kernel, and then it is convolved with the input feature maps to produce the output feature maps.  ... 
arXiv:2007.12928v3 fatcat:nxoejcfdnzas3jznbqsale36ty

Learning Guided Convolutional Network for Depth Completion [article]

Jie Tang, Fei-Peng Tian, Wei Feng, Jian Li, Ping Tan
2019 arXiv   pre-print
Inspired by the guided image filtering, we design a novel guided network to predict kernel weights from the guidance image.  ...  Dense depth perception is critical for autonomous driving and other robotics applications. However, modern LiDAR sensors only provide sparse depth measurement.  ...  Depth-only Methods These methods use a sparse or lowresolution depth image as input to generate a full-resolution depth map.  ... 
arXiv:1908.01238v1 fatcat:n2k6zuoxtbdudkartuxquepqsu

Single-Image Super-Resolution Neural Network via Hybrid Multi-Scale Features

Wenfeng Huang, Xiangyun Liao, Lei Zhu, Mingqiang Wei, Qiong Wang
2022 Mathematics  
In this paper, we propose an end-to-end single-image super-resolution neural network by leveraging hybrid multi-scale features of images.  ...  Experiments on five popular benchmarks demonstrate that our super-resolution approach achieves better performance with fewer parameters and less memory consumption, compared to more than 20 SOTAs.  ...  Unlike L2, the L1 loss function is widely used in many super-resolution tasks; many experiments have shown that it improves the effect of super-resolution.  ... 
doi:10.3390/math10040653 fatcat:6u2u6b2x5vb6rmc57wbjgnc46e

Deep Convolutional Grid Warping Network for Joint Depth Map Upsampling

Yoonmo Yang, Dongsin Kim, Byung Tae Oh
2020 IEEE Access  
In this paper, we propose a novel approach to upsample depth maps by using geometric deformation instead of pixel value refinement, which is employed in a majority of existing methods.  ...  Depth maps play an important role in the representation of 3D information.  ...  [17] focused on the receptive field for depth map upsampling; the receptive fields were enlarged using deformable kernel convolution (DKN).  ... 
doi:10.1109/access.2020.3015209 fatcat:s7minlrtuvcp7mqxu2sldetrgm
« Previous Showing results 1 — 15 out of 2,173 results