A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2018; you can also visit the original URL.
The file type is application/pdf.
Filters
Task-Aware Image Downscaling
[chapter]
2018
Lecture Notes in Computer Science
Preserved Fulltext
We propose an auto-encoder-based framework that enables joint learning of the downscaling network and the upscaling network to maximize the restoration performance. ...
In this paper, we present a novel technique called task-aware image downscaling to support an upscaling task. ...
Figure 1 shows an example of the importance of choosing an appropriate downscaling method, where the downscaled LR images in blue and red look similar, but the restored HR image Fig. 1 : Our task-aware ...
doi:10.1007/978-3-030-01225-0_25
fatcat:3j3pgs5zznd77mgfeer34oalc4
HyperThumbnail: Real-time 6K Image Rescaling with Rate-distortion Optimization
[article]
2023
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2023; you can also visit the original URL.
The file type is application/pdf.
Then, an efficient frequency-aware decoder reconstructs a high-fidelity HR image from the LR one in real time. ...
To address these two challenges, we propose a novel framework (HyperThumbnail) for real-time 6K rate-distortion-aware image rescaling. ...
In this work, we propose the HyperThumbnail, a ratedistortion-aware framework for 6K real-time image rescaling, as shown in Table 1 (c). ...
arXiv:2304.01064v1
fatcat:tp6zyl7spffpljhoikejd4frcy
Effective Invertible Arbitrary Image Rescaling
[article]
2022
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2022; you can also visit the original URL.
The file type is application/pdf.
To increase its real world applicability, numerous models have also been proposed to restore SR images with arbitrary scale factors, including asymmetric ones where images are resized to different scales ...
(INN) are able to increase upscaling accuracy significantly by optimizing the downscaling and upscaling cycle jointly. ...
As illustrated in Table 5 , three most relevant models which are optimized jointly for downscaling and upscaling are included so the inference time includes both downscaling and upscaling and listed separately ...
arXiv:2209.13055v1
fatcat:a4xqwkrisrbrjormohe5bseetm
Scale-arbitrary Invertible Image Downscaling
[article]
2022
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2022; you can also visit the original URL.
The file type is application/pdf.
Other Versions
Recent invertible image downscaling methods jointly model the downscaling/upscaling problems and achieve significant improvements. ...
Meanwhile, the HR information is embedded in the downscaled low-resolution (LR) counterparts in a nearly imperceptible form such that our AIDN can also restore the original HR images solely from the LR ...
Unlike the scale-aware upsampling layer, our CRM can be used for both downscaling and upscaling, and is not only scale-aware but also content-adaptive. ...
arXiv:2201.12576v3
fatcat:64lnvqg7fbabxddemsn5lqz43m
VisionISP: Repurposing the Image Signal Processor for Computer Vision Applications
2019
2019 IEEE International Conference on Image Processing (ICIP)
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL.
The file type is application/pdf.
Note that this fulltext copy is not of the "primary" version of this work. The version it corresponds to is:
2019 pre-print arXiv:1911.05931v1 fatcat:apgwyjmdvbf7fioemz67iimmwe
The blocks in VisionISP are simple, content-aware, and trainable. ...
However, optimal perceptual image quality does not always translate into optimal performance for computer vision applications. ...
For an imaging system designed with real-time computer vision applications in mind, the connection between an ISP and a computer vision engine would be as important as the processing blocks within these ...
doi:10.1109/icip.2019.8803607
dblp:conf/icip/WuISNGSAM19
fatcat:ze5j3i2mwvdenmuc3h5op2ldla
Invertible Rescaling Network and Its Extensions
[article]
2022
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2022; you can also visit the original URL.
The file type is application/pdf.
downscaled and decolorized images, and rate-distortion of image compression. ...
Image rescaling is a commonly used bidirectional operation, which first downscales high-resolution images to fit various display screens or to be storage- and bandwidth-friendly, and afterward upscales ...
The authors would like to thank Yaolong Wang, Di He, Guolin Ke and Jiang Bian for their help on discussions, experiments and writing in the preliminary version of this paper. ...
arXiv:2210.04188v1
fatcat:rddzamq4m5bjlpqzqqmpizgdki
CaDM: Codec-aware Diffusion Modeling for Neural-enhanced Video Streaming
[article]
2023
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2023; you can also visit the original URL.
The file type is application/pdf.
Other Versions
the enhancement on the decoder side while omitting the co-design of encoder, (2) limited generative capacity to recover high-fidelity perceptual details, and (3) optimizing the compression-and-restoration ...
Second, CaDM empowers the decoder with high-quality enhancement by making the denoising diffusion restoration aware of encoder's resolution-color conditions. ...
., TecoGAN [8] , PULSE [42] and Real-ESRGAN [59] . ...
arXiv:2211.08428v2
fatcat:vohiw7r35fbx3k26nwuflhfiqm
Real-World Image Super-Resolution by Exclusionary Dual-Learning
[article]
2022
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2022; you can also visit the original URL.
The file type is application/pdf.
Although deep learning-based methods have achieved promising restoration quality on real-world image super-resolution datasets, they ignore the relationship between L1- and perceptual- minimization and ...
In this paper, we discuss the image types within a corrupted image and the property of perceptual- and Euclidean- based evaluation protocols. ...
Wei et al. proposed a DASR framework [24] by calculating the domain distance between LR images and real images with the domain-gap aware training and domaindistance weighted supervision. Ji et al. ...
arXiv:2206.02609v1
fatcat:66dwtmzd5vdt5kzuo7sk7wa2xu
Deep Learning-based Image Super-Resolution Considering Quantitative and Perceptual Quality
[article]
2019
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL.
The file type is application/pdf.
Other Versions
the ground-truth images and the naturalness, respectively. ...
Recently, it has been shown that in super-resolution, there exists a tradeoff relationship between the quantitative and perceptual quality of super-resolved images, which correspond to the similarity to ...
times through the ×2 path, and by passing through the ×8 path and then downscaling via bicubic interpolation. ...
arXiv:1809.04789v2
fatcat:3r423fq6szbg3mrhx4orwf2hri
NTIRE 2019 Challenge on Image Enhancement: Methods and Results
2019
2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW)
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2022; you can also visit the original URL.
The file type is application/pdf.
This paper reviews the first NTIRE challenge on perceptual image enhancement with the focus on proposed solutions and results. ...
The target metric used in this challenge combined fidelity scores (PSNR and SSIM) with solutions' perceptual results measured in a user study. ...
Introduction Image restoration and image enhancement are among the fundamental computer vision problems aiming at the improvement of different image quality aspects, including its perceptual quality, resolution ...
doi:10.1109/cvprw.2019.00275
dblp:conf/cvpr/IgnatovT19
fatcat:nejsyrgv3raodaglc5zzaysamy
Progressive Semantic-Aware Style Transformation for Blind Face Restoration
[article]
2021
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2021; you can also visit the original URL.
The file type is application/pdf.
Other Versions
Face restoration is important in face image processing, and has been widely studied in recent years. ...
Specifically, instead of using an encoder-decoder framework as previous methods, we formulate the restoration of LQ face images as a multi-scale progressive restoration procedure through semantic-aware ...
This work was partially supported by Alibaba DAMO Academy, Hong Kong RGC RIF grant (R5001-18), and Hong Kong RGC GRF grant (project# 17203119). ...
arXiv:2009.08709v2
fatcat:n7wutdi7bjhszhibbyo5xidate
Pan-Sharpening with Color-Aware Perceptual Loss and Guided Re-Colorization
[article]
2020
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL.
The file type is application/pdf.
We present a novel color-aware perceptual (CAP) loss for learning the task of pan-sharpening. ...
Additionally, we propose "guided re-colorization", which generates a pan-sharpened image with real colors from the MS input by "picking" the closest MS pixel color for each pan-sharpened pixel, as a human ...
Perceptual loss has shown to be more effective for image restoration than the plain l1 or l2 losses. ...
arXiv:2006.16583v1
fatcat:vvjleecugrbh3ipxugqsrocq3u
Joint Learning of Super-Resolution and Perceptual Image Enhancement for Single Image
2021
IEEE Access
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2021; you can also visit the original URL.
The file type is application/pdf.
JOINT SR-PIE In real-world application, it is impossible to remain image perceptual quality when resolving SR problem. ...
Whereas, the requirement of abundant parameters limits its application in real-time practice. ...
doi:10.1109/access.2021.3068861
fatcat:el7xciykonbihhmpel3vge3uqy
AIM 2020 Challenge on Learned Image Signal Processing Pipeline
[article]
2020
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL.
The file type is application/pdf.
The target metric used in this challenge combined fidelity scores (PSNR and SSIM) with solutions' perceptual results measured in a user study. ...
The participating teams were solving a real-world RAW-to-RGB mapping problem, where to goal was to map the original low-quality RAW images captured by the Huawei P20 device to the same photos obtained ...
Acknowledgments We thank the AIM 2020 sponsors: Huawei, MediaTek, Qualcomm, NVIDIA, Google and Computer Vision Lab / ETH Zürich. ...
arXiv:2011.04994v1
fatcat:i43uj46xxfhcxgsg7e7l6qbvqe
Best-Buddy GANs for Highly Detailed Image Super-Resolution
[article]
2021
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2021; you can also visit the original URL.
The file type is application/pdf.
Note that this fulltext copy is not of the "primary" version of this work. The version it corresponds to is:
2021 pre-print arXiv:2103.15295v2 fatcat:yg355fm3zrgunow7nr6exdhcvuOther Versions
We consider the single image super-resolution (SISR) problem, where a high-resolution (HR) image is generated based on a low-resolution (LR) input. ...
Besides, we propose a region-aware adversarial learning strategy that directs our model to focus on generating details for textured areas adaptively. ...
is trained to distinguish between recovered results and real natural images. ...
arXiv:2103.15295v3
fatcat:xljujqfacjdevnmgn5vdrdq764
« Previous
Showing results 1 — 15 out of 275 results