AIM 2020 Challenge on Learned Image Signal Processing Pipeline

Ignatov, Andrey; Timofte, Radu; Zhang, Zhilu; Liu, Ming; Wang, Haolin; Zuo, Wangmeng; Zhang, Jiawei; Zhang, Ruimao; Peng, Zhanglin; Ren, Sijie; Dai, Linhui; Liu, Xiaohong; Li, Chengqi; Chen, Jun; Itô, Yûichi; Vasudeva, Bhavya; Deora, Puneesh; Pal, Umapada; Guo, Zhenyu; Zhu, Yu; Liang, Tian; Li, Chenghua; Leng, Cong; Pan, Zhihong; Li, Baopu; Kim, Byung–Hoon; Song, Joonyoung; Ye, Jong Chul; Baek, Jehyun; Zhussip, Magauiya; Koishekenov, Yeskendir; Ye, Hwechul Cho; Liu, Xin; Hu, Xueying; Jiang, Jun; Gu, Jun; Li, Kai; Tan, Pengliang; Hou, Bingxin

doi:10.1007/978-3-030-67070-2_9

Cited by 49 publications

(23 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This data-set was featured in the ECCV 2020 AIM Learned Smartphone ISP Challenge [18]. At the time of writing we could find sufficient detail on [6] and [21] publicly available for a comparison of single network fidelity and computational cost.…”

Section: Data-setsmentioning

confidence: 99%

Overparametrization of HyperNetworks at Fixed FLOP-Count Enables Fast Neural Image Enhancement

Müller¹

2021

Preprint

View full text Add to dashboard Cite

Deep convolutional neural networks can enhance images taken with small mobile camera sensors and excel at tasks like demoisaicing, denoising and super-resolution. However, for practical use on mobile devices these networks often require too many FLOPs and reducing the FLOPs of a convolution layer, also reduces its parameter count. This is problematic in view of the recent finding that heavily overparameterized neural networks are often the ones that generalize best.In this paper we propose to use HyperNetworks to break the fixed ratio of FLOPs to parameters of standard convolutions. This allows us to exceed previous state-of-the-art architectures in SSIM and MS-SSIM on the Zurich RAWto-DSLR (ZRR) data-set at > 10× reduced FLOP-count. On ZRR we further observe generalization curves consistent with 'double-descent' behavior at fixed FLOP-count, in the large image limit. Finally we demonstrate the same technique can be applied to an existing network (VDN) to reduce its computational cost while maintaining fidelity on the Smartphone Image Denoising Dataset (SIDD).

show abstract

Section: Data-setsmentioning

confidence: 99%

Overparametrization of HyperNetworks at Fixed FLOP-Count Enables Fast Neural Image Enhancement

Müller¹

2021

Preprint

View full text Add to dashboard Cite

show abstract

“…However, the model could not correctly reconstruct the color of the RGB images in many cases. We report the results from the AIM 2020 learned smartphone ISP challenge [6]. There were two separate tracks which we both participated in.…”

Section: Network Trainingmentioning

confidence: 99%

PyNET-CA: Enhanced PyNET with Channel Attention for End-to-End Mobile Image Signal Processing

Kim

Song

et al. 2020

Computer Vision – ECCV 2020 Workshops

Self Cite

View full text Add to dashboard Cite

Reconstructing RGB image from RAW data obtained with a mobile device is related to a number of image signal processing (ISP) tasks, such as demosaicing, denoising, etc. Deep neural networks have shown promising results over hand-crafted ISP algorithms on solving these tasks separately, or even replacing the whole reconstruction process with one model. Here, we propose PyNET-CA, an end-to-end mobile ISP deep learning algorithm for RAW to RGB reconstruction. The model enhances PyNET, a recently proposed state-of-the-art model for mobile ISP, and improve its performance with channel attention and subpixel reconstruction module. We demonstrate the performance of the proposed method with comparative experiments and results from the AIM 2020 learned smartphone ISP challenge. The source code of our implementation is available at https://github.com/egyptdj/skyb-aim2020-public

show abstract

“…This challenge is one of the AIM 2020 associated challenges on: scene relighting and illumination estimation [8], image extreme inpainting [31], learned image signal processing pipeline [15], rendering realistic bokeh [16], real image super-resolution [37], efficient super-resolution [42], and video extreme superresolution [9]. Our development phase has started on May 1st, and the test phase is opened on July 10th.…”

Section: Aim 2020 Vtsr Challengementioning

confidence: 99%

AIM 2020 Challenge on Video Temporal Super-Resolution

Son¹,

Lee²,

Nah³

et al. 2020

Preprint

Self Cite

View full text Add to dashboard Cite

Videos in the real-world contain various dynamics and motions that may look unnaturally discontinuous in time when the recorded frame rate is low. This paper reports the second AIM challenge on Video Temporal Super-Resolution (VTSR), a.k.a. frame interpolation, with a focus on the proposed solutions, results, and analysis. From low-framerate (15 fps) videos, the challenge participants are required to submit higher-frame-rate (30 and 60 fps) sequences by estimating temporally intermediate frames. To simulate realistic and challenging dynamics in the real-world, we employ the REDS VTSR dataset derived from diverse videos captured in a hand-held camera for training and evaluation purposes. There have been 68 registered participants in the competition, and 5 teams (one withdrawn) have competed in the final testing phase. The winning team proposes the enhanced quadratic video interpolation method and achieves state-of-the-art on the VTSR task.

show abstract

AIM 2020 Challenge on Learned Image Signal Processing Pipeline

Cited by 49 publications

References 30 publications

Overparametrization of HyperNetworks at Fixed FLOP-Count Enables Fast Neural Image Enhancement

Overparametrization of HyperNetworks at Fixed FLOP-Count Enables Fast Neural Image Enhancement

PyNET-CA: Enhanced PyNET with Channel Attention for End-to-End Mobile Image Signal Processing

AIM 2020 Challenge on Video Temporal Super-Resolution

Contact Info

Product

Resources

About