• Home
  • ETH Zurich
  • Department Information Technology and Electrical Engineering
  • Yawei Li
Yawei Li

Yawei Li
ETH Zurich | ETH Zürich · Department Information Technology and Electrical Engineering

Doctor of Philosophy

About

42
Publications
7,705
Reads
How we measure 'reads'
A 'read' is counted each time someone views a publication summary (such as the title, abstract, and list of authors), clicks on a figure, or views or downloads the full-text. Learn more
855
Citations
Citations since 2016
42 Research Items
856 Citations
20162017201820192020202120220100200300400
20162017201820192020202120220100200300400
20162017201820192020202120220100200300400
20162017201820192020202120220100200300400
Introduction
My research focuses on developing efficient models for computer vision tasks including high-level vision (image classification, visual tracking), low-level vision (image super-resolution, denoising), and point cloud processing (graph neural networks). To achieve that, one way is to design efficient models directly. The other direction is to apply network compression algorithms to predefined networks. And my research proceeds in both directions.
Additional affiliations
September 2017 - present
ETH Zurich
Position
  • PhD Student
Education
September 2014 - June 2017
University of Electronic Science and Technology of China
Field of study
  • Communication and Information System

Publications

Publications (42)
Chapter
Reference-based image super-resolution (RefSR) aims to exploit auxiliary reference (Ref) images to super-resolve low-resolution (LR) images. Recently, RefSR has been attracting great attention as it provides an alternative way to surpass single image SR. However, addressing the RefSR problem has two critical challenges: (i) It is difficult to match...
Preprint
Reference-based image super-resolution (RefSR) aims to exploit auxiliary reference (Ref) images to super-resolve low-resolution (LR) images. Recently, RefSR has been attracting great attention as it provides an alternative way to surpass single image SR. However, addressing the RefSR problem has two critical challenges: (i) It is difficult to match...
Preprint
Full-text available
Channel (or 3D filter) pruning serves as an effective way to accelerate the inference of neural networks. There has been a flurry of algorithms that try to solve this practical problem, each being claimed effective in some ways. Yet, a benchmark to compare those algorithms directly is lacking, mainly due to the complexity of the algorithms and some...
Preprint
Full-text available
This paper reviews the NTIRE 2022 challenge on efficient single image super-resolution with focus on the proposed solutions and results. The task of the challenge was to super-resolve an input image with a magnification factor of $\times$4 based on pairs of low and corresponding high resolution images. The aim was to design a network for single ima...
Preprint
While recent years have witnessed a dramatic upsurge of exploiting deep neural networks toward solving image denoising, existing methods mostly rely on simple noise assumptions, such as additive white Gaussian noise (AWGN), JPEG compression noise and camera sensor noise, and a general-purpose blind denoising method for real images remains unsolved....
Preprint
Full-text available
Video restoration (e.g., video super-resolution) aims to restore high-quality frames from low-quality frames. Different from single image restoration, video restoration generally requires to utilize temporal information from multiple adjacent but usually misaligned video frames. Existing deep methods generally tackle with this by exploiting a slidi...
Article
Recent works on plug-and-play image restoration have shown that a denoiser can implicitly serve as the image prior for model-based methods to solve many inverse problems. Such a property induces considerable advantages for plug-and-play image restoration (e.g., integrating the flexibility of model-based method and effectiveness of learning-based me...
Preprint
Video super-resolution (VSR), with the aim to restore a high-resolution video from its corresponding low-resolution version, is a spatial-temporal sequence prediction problem. Recently, Transformer has been gaining popularity due to its parallel computing ability for sequence-to-sequence modeling. Thus, it seems to be straightforward to apply the v...
Preprint
Full-text available
In this paper, we aim at improving the computational efficiency of graph convolutional networks (GCNs) for learning on point clouds. The basic graph convolution that is typically composed of a $K$-nearest neighbor (KNN) search and a multilayer perceptron (MLP) is examined. By mathematically analyzing the operations there, two findings to improve th...
Preprint
Full-text available
We study how to introduce locality mechanisms into vision transformers. The transformer network originates from machine translation and is particularly good at modelling long-range dependencies within a long sequence. Although the global interaction between the token embeddings could be well modelled by the self-attention mechanism of transformers,...
Preprint
Open compound domain adaptation (OCDA) is a domain adaptation setting, where target domain is modeled as a compound of multiple unknown homogeneous domains, which brings the advantage of improved generalization to unseen domains. In this work, we propose a principled meta-learning based approach to OCDA for semantic segmentation, MOCDA, by modeling...
Chapter
Network pruning has been the driving force for the acceleration of neural networks and the alleviation of model storage/transmission burden. With the advent of AutoML and neural architecture search (NAS), pruning has become topical with automatic mechanism and searching based architecture optimization. Yet, current automatic designs rely on either...
Preprint
Full-text available
This paper reviews the AIM 2020 challenge on efficient single image super-resolution with focus on the proposed solutions and results. The challenge task was to super-resolve an input image with a magnification factor x4 based on a set of prior examples of low and corresponding high resolution images. The goal is to devise a network that reduces on...
Preprint
Full-text available
Recent works on plug-and-play image restoration have shown that a denoiser can implicitly serve as the image prior for model-based methods to solve many inverse problems. Such a property induces considerable advantages for plug-and-play image restoration (e.g., integrating the flexibility of model-based method and effectiveness of learning-based me...
Preprint
Full-text available
In this paper, we tackle the problem of convolutional neural network design. Instead of focusing on the overall architecture design, we investigate a design space that is usually overlooked, \ie adjusting the channel configurations of predefined networks. We find that this adjustment can be achieved by pruning widened baseline networks and leads to...
Preprint
Full-text available
These days, unsupervised super-resolution (SR) has been soaring due to its practical and promising potential in real scenarios. The philosophy of off-the-shelf approaches lies in the augmentation of unpaired data, i.e. first generating synthetic low-resolution (LR) images $\mathcal{Y}^g$ corresponding to real-world high-resolution (HR) images $\mat...
Preprint
Full-text available
Network pruning has been the driving force for the efficient inference of neural networks and the alleviation of model storage and transmission burden. Traditional network pruning methods focus on the per-filter influence on the network accuracy by analyzing the filter distribution. With the advent of AutoML and neural architecture search (NAS), pr...
Preprint
Full-text available
In this paper, we analyze two popular network compression techniques, i.e. filter pruning and low-rank decomposition, in a unified sense. By simply changing the way the sparsity regularization is enforced, filter pruning and low-rank decomposition can be derived accordingly. This provides another flexible choice for network compression because the...
Chapter
This paper reviews the AIM 2020 challenge on efficient single image super-resolution with focus on the proposed solutions and results. The challenge task was to super-resolve an input image with a magnification factor \(\times \)4 based on a set of prior examples of low and corresponding high resolution images. The goal is to devise a network that...
Preprint
Full-text available
Convolutional neural networks (CNNs) based solutions have achieved state-of-the-art performances for many computer vision tasks, including classification and super-resolution of images. Usually the success of these methods comes with a cost of millions of parameters due to stacking deep convolutional layers. Moreover, quite a large number of filter...
Preprint
Full-text available
We tackle the problem of retrieving high-resolution (HR) texture maps of objects that are captured from multiple view points. In the multi-view case, model-based super-resolution (SR) methods have been recently proved to recover high quality texture maps. On the other hand, the advent of deep learning-based methods has already a significant impact...
Chapter
This paper reviews the first challenge on efficient perceptual image enhancement with the focus on deploying deep learning models on smartphones. The challenge consisted of two tracks. In the first one, participants were solving the classical image super-resolution problem with a bicubic downscaling factor of 4. The second track was aimed at real-w...
Chapter
Full-text available
Although the accuracy of super-resolution (SR) methods based on convolutional neural networks (CNN) soars high, the complexity and computation also explode with the increased depth and width of the network. Thus, we propose the convolutional anchored regression network (CARN) for fast and accurate single image super-resolution (SISR). Inspired by l...
Preprint
Full-text available
This paper reviews the first challenge on efficient perceptual image enhancement with the focus on deploying deep learning models on smartphones. The challenge consisted of two tracks. In the first one, participants were solving the classical image super-resolution problem with a bicubic downscaling factor of 4. The second track was aimed at real-w...
Article
Hybrid videos that contain periodic low-resolution (LR) frames and high-resolution (HR) guide frames are largely used for the consideration of bandwidth efficiency and the tradeoff between spatial and temporal resolution. Super-resolution (SR) algorithms are necessary to refine the LR frames, in which non-local means (NLM) is a promising algorithm....
Conference Paper
Power consumption, transmission bandwidth, and spatial-temporal resolution tradeoff are among the most important factors that affect video processing on mobile devices. Scalable video coding (SVC) provides a possible solution to overcome these problems. Every video in SVC format consists of high-resolution (HR) frames and low-resolution (LR) frames...
Article
Full-text available
Visual restoration and recognition are traditionally addressed in pipeline fashion, i.e. denoising followed by classification. Instead, observing correlations between the two tasks, for example clearer image will lead to better categorization and vice visa, we propose a joint framework for visual restoration and recognition for handwritten images,...
Article
Super-resolution (SR) techniques, which are based on single or multi-frame low-resolution (LR) images, have been extensively investigated in the last two decades. Mixed-resolution multiview video format plays an important role in three-dimensional television (3DTV) coding scheme. Previous work considers multiview or multi-camera images and videos a...
Conference Paper
Full-text available
Multiview video super-resolution provides a promising solution to the contradiction between the huge data size of multiview video and the degraded video quality due to mixed-resolution compression. This algorithm consists of two different functional layers. An information extraction layer draws relevant high-frequency information from the high-reso...
Article
Full-text available
Non-Gaussian quadratures, especially Cauchy quadratures, are widely used to model impulsive noise environments. The envelope and phase statistics of Cauchy quadratures are investigated. The probability density functions and cumulative distribution functions of the envelope and phase of Cauchy quadratures are given in closed-form expression. The cha...

Network

Cited By