About the Lab

The Computer Vision Laboratory, ETH Zurich, works on the computer-based interpretation of 2D and 3D image data sets from conventional and non-conventional image sources.

The computer vision lab performs research in the fields of Medical Image Analysis and Visualization, Object Recognition, Gesture Analysis, Tracking, and Scene Understanding and Modeling

We treat the complete cycle from signals to their interpretation, and to the resulting action. It is our objective to develop universal concepts and methods. In order to meet these challenges, we want to keep our finger at the pulse of international, ongoing research. Therefore, we operate in the context of many collaborations with other labs, like the CO-ME and IM2 Swiss competence centers (NCCR), and many European projects.

This said, it simultaneously is our strategy to let difficult, real-world applications drive our research and development. We see collaboration with industry as an important plus for an engineering lab like ours.

The groups in CVL are led by Prof. Luc Van GoolProf. Gabor Székely, Prof. Ender Konukoglu and Prof. Orçun Göksel.

Vision Lab News

Results of the joint WSL/ETH research project "Automated detection of invasive alien plants along highways" have been published in 20 min and SRF1 (min 2:15).

"Computers are sometimes better at mapping plant species than humans" (DIAGONAL - WSL).
Our group developed and adapted artificial intelligence (deep learning) algorithms to work with images and video streams captured along Swiss highways from cameras mounted on cars.

More info at:



20 min


Best Student Paper Award at BMVC 2019:
The work by Ardhendu ShekharTripathi, Martin Danelljan, Luc Van Gool and Radu Timofte titled "Tracking the Known and the Unknown by Leveraging Semantic Information" won the best student paper award the 30th British Machine Vision Conference 2019 held in Cardiff, UK. (bmvc link: https://bmvc2019.org )

Current research in visual tracking is largely focused on the generic case, where no prior knowledge about the target object is assumed. However, many real-world tracking applications stem from specific scenarios where the class or type of object is known. In this work, is proposed a tracking framework that can exploit this semantic information, without sacrificing the generic nature of the tracker. In addition to the target-specific appearance, the class of the object is modeled through a semantic module that provides complementary class-specific predictions.  By further integrating a semantic classification module, the learned class-specific models can be utilized even if the target class is unknown. The proposed unified tracking architecture is trained end-to-end on large scale tracking datasets by exploiting the available semantic metadata. Comprehensive experiments are performed on five tracking benchmarks. The proposed approach achieves state-of-the-art performance while operating at real-time frame-rates.

Yawei Li received the best poster presentation prize at the International Computer Vision Summer School ICVSS 2019.
The work "3D Appearance Super-Resolution with Deep Learning" (Y.Li, V.Tsiminaki, R.Timofte, M.Pollefeys, L.Van Gool), was previously published at CVPR 2019. This is the first deep learning-based approach to super-resolve the appearance (texture maps) of 3D objects.  Also, a 3D appearance super-resolution (3DASR) dataset of low and high res textures pairs of 3D objects is proposed.


Simon Hecker, Henrik Maurenbrecher, Alex Liniger, Dengxin Dai and Luc Van Gool published the ITSC 2019 work "Learning a Curve Guardian for Motorcycles" which caught great attention in different News outlets.
Download paper

Venturebeat: "... Turns taken too quickly or sharply are responsible for nearly a fifth of all motorcycle accidents, according to estimates, which also suggest misjudged curves are to blame for 15% of rider fatalities. Fortunately, researchers at ETH Zurich and KU Leuven have recently proposed a solution in a paper on the preprint server Arxiv.org (“Learning a Curve Guardian for Motorcycles“). Their road curvature warning system leverages computer vision and mapping data to predict lane position and motorcycle roll angle, in addition to the road geometry of future paths."

RideApart: " ... It’s fascinating work, and although this paper currently exists only in a pre-published state, it will be fascinating to see what real-world developments come from this research in the future."

Moto.it: Simon Hecker, Alexander Liniger, Henrik Maurenbrecher, Dengxin Dai e Luc Van Gool, dello ETH di Zurigo e del KU di Lovanio, hanno pubblicato un interessante studio che va ad arricchire il dibattito sulla necessità di sistemi Advanced Rider Assistance Systems (ARAS) legati all'intelligenza artificiale che possano aiutare il motociclista ad evitare questo genere di incidenti: come?


2nd place in the D2-City & BDD100K Detection Domain Adaptation Challenge for autonomous driving
A team composed of members of ETH-CVL (Yuhua Chen, Wen Li, and Luc Van Gool), and the University of Electronic Science and Technology of China (Qing Lian and Lixin Duan) have gained 2nd place in the D2-City & BDD100K Detection Domain Adaptation Challenge for autonomous driving, at the IEEE Conf. on Computer Vision and Pattern Recognition 2019.


Article published in the World Economic Forum Newsletter presenting CVL- work related to turning night images into day images (published by Julien Weissenberg). Please visit the Website of the World Economic Forum for more information.


CVL-paper selected as most interesting publication of the week by DeepAI (20.04.2019):
Sliced Wasserstein Generative Models (CVPR 2019). Jiqing Wu, Zhiwu Huang, Dinesh Acharya, Wen Li, Janine Thoma, Danda Paudel, and Luc Van Gool.
Link to DeepAI Website

“Hochschulmedizin Zurich Flagship Project is awarded to “SURGENT: Surgeon Enhancing Technologies” with an inter-institutional consortium including D-ITET professors Ender Konukoglu and Orcun Goksel.
HMZ Surgent  <https://www.hochschulmedizin.uzh.ch/de/projekte/surgent/konsortium.html>
ETH Zurich news  <https://www.ethz.ch/en/news-and-events/eth-news/news/2018/11/surgent-hochschulmedizin-zurich.html>

Prof. Luc Van Gool receives Distinguished Researcher award
At its International Conference on Computer Vision 2017 in Venice, the IEEE Computer Society granted prof. Luc Van Gool the Distinguished Researcher award.


Dr. Sanabria and Prof. Goksel of CVL receive the 2017 Venture Best Business Idea
Congratulations to Sergio Sanabria and Orcun Goksel of the Computer Vision Lab at ETH Zurich for receiving the Venture Best Business Idea award with ReflectUS, among over 200 applicants around Switzerland. The award was announced at a ceremony with the federal Councilor and several company representatives.


The ERC project VarCity goes to the cinema!
The project VarCity was premiered as a documentary movie in a cinema in Zurich. The premiere as well as the project itself received international news coverage in more than 100 reports and broadcasts.


A Google Faculty Research Award goes to CVL
Prof. Van Gool and dr. Dengxin Dai received a Google Faculty Research Award for Question Answering with Textual and Visual Answers. The selection process was quite selective, as only 143 awards were granted out of 876 proposals, submitted by over 300 universities from 44 countries.


CVL receives Koenderink Prize
At the European Conference on Computer Vision 2016 in Amsterdam CVL's paper `SURF: speeded-up robust features' by H. Bay, T. Tuytelaars, and L. Van Gool was rewarded the prestigious Koenderink Prize, given to `fundamental contributions in computer vision that stood the test of time'.


Dengxin Dai's research proposal selected as finalist of Qualcomm Innovation Fellowship
Dengxin Dai's research proposal, Look&Tell: Voice-user Interface to Pictures, has been selected as one of the finalist proposals for the Qualcomm Innovation Fellowship 2016. Dengxin will present his proposal at the University of Amsterdam on May 10th 2016, in order to make it for the final selection.


Spark Award 2016 goes to the CAiM group in CVL
Prof. Göksel's group of Computer-assisted Applications in Medicine (CAiM) in CVL has received the Spark Award 2016 for the most promising ETH invention of the year. This was awarded for the invention ``Hand-held ultrasound imaging of speed-of-sound for breast examination'' of Sergio Sanabria and Orcun Goksel.


faces.ethz.ch goes viral around the world - 1 million users in 12 hours after launch
Recently CVL released faces.ethz.ch, a website where people can upload a photo of their face and an algorithm will then estimate the age, gender and facial attractiveness of the person. Within a couple of hours the website gained large popularity around the world with more than 1 million people accessing the website in the first 12 hours. The software uses Deep Neural Networks which were trained on millions of images. The website was an outcome of the research conducted by Rasmus RotheRadu Timofte and Luc van Gool in collaboration with the Swiss dating startup Blinq and resulted in the paper  Some like it hot - visual guidance for preference prediction.


CVL wins apparent age estimation challenge and best paper award at ICCV 2015 workshop
Rasmus Rothe, Radu Timofte and Luc van Gool win the apparent age estimation challenge with more than 100 participating teams at the ChaLearn Looking at People Workshop at ICCV 2015. The goal is to predict the apparent age of people from RGB images in-the-wild. Their paper DEX: Deep EXpectation of apparent age from a single image also won the NVIDIA ChaLearn LAP 2015 Best Paper Award.


CVL spin-off Face-Shift bought by Apple
Computer Vision Lab spin-off Face-Shift has been bought by Apple. Face-Shift offers groundbreaking technology for the capturing of facial expressions, and their subsequent real-time transfer to avatars.


CVL spin-off Parquery featured in the news and on ERC Stories
After having launched the first computer vision-based smart parking project in Locarno, Switzerland, Parquery has been featured on the Swiss Italian news. It has also been showcased by the European Research Council with an ERC Story, presenting Parquery and Varcity together.


CVL spin-off uniqFEED launched
Mid 2015 spin-off uniqFEED got launched, after a successful first round of capital raising. The company has already hired its first collaborators, but is looking for more! uniqFEED specializes in Augmented Reality applications for the sports sector. It recently got additional funding through a Proof-of-Concept project connected to CVL's advanced ERC grant VarCity (Variation & the City, about 4D city modeling).


CVL receives 3rd place in Philips/IPCAI Best Paper Award
The work "Thermal Expansion Imaging for Monitoring Lesion Depth Using M-Mode Ultrasound During Cardiac RF Ablation: In-Vitro Study" by Peter Baki, Sergio J Sanabria, Gabor Kosa, Gabor Szekely, and Orcun Goksel has received the 3rd place in Philips/IPCAI Best Paper Award, out of 100 submissions. IPCAI is a focused conference on Information Processing in Computer-Assisted Interventions that took place this year in Barcelona, Spain. The awarded work, accordingly, also featured in a special issue of the International Journal of Computer Assisted Radiology and Surgery.


CVL spin-off upicto acquired
In 2014, CVL spin-off upicto has been acquired by Logitech. The spin-off has specialised in advanced video interpretation. Logitech created a special research office in Zurich to host the former upicto team, and to explore future horizons while staying close to ETH.


CVL receives the Best Poster Award at 4D Treatment Planning Workshop
The work title "Total Variation Regularization of Displacements in Parametric Image Registration" by Valeriy Vishnevskiy, Tobias Gass, Gabor Szekely, and Orcun Goksel has received the Best Poster Award at the 4D Treatment Planning Workshop, London, UK, November 2014. This is an annual event focusing on planning and delivery of radiotherapy to moving targets, for both photon and particle therapies.


CTI MedTech Award 2014 goes to CVL team 
Gábor Székely, Head of CTI (Commission for Technology and Innovation) Medtech, presented Orçun Göksel, SNSF-Professor at the Department of Information Technology and Electrical Engineering at ETH Zurich and Vincent Baumann (Veinpress GmbH) with the award, which includes a check for more than CHF 10 000 and a trophy. Their project was selected from among three nominated projects by the audience of around 400 experts from the fields of Medtech research and industry. 
Read more


ETH Robocup team powered by CVL 
Several enthusiastic students, the Computer Vision Lab and the Automatic Control Lab (both D-ITET) have sent their robot soccer team to Magdeburg early April 2014. In the Standard Platform League at the Robocup German Open tournament, the 4 robots have defended ETH's honor with great success. 
Read more (in German)


CVL spin-off kooaba acquired by Qualcomm 
The CVL spin-off kooaba has been created in 2007 by two former PhD students, Till Quack and Herbert Bay. It brought visual recognition to the mobile phone, thus connecting the real and the digital worlds. Early 2013 the company has been acquired by Qualcomm, one of the world's biggest electronic companies. Qualcomm uses the kooaba technology to further strengthen in Vuforia cloud-based visual recognition and augmented reality tool. 
Read more


CVL in the Swiss TV 
An interview with Ph.D. Matthias Harders is featured by Swiss TV in the "Einstein" show (June 6th, 2013). 
Read more


An article on the CVL members participating in RoboCup 2013 
Two CVL members, Gemma Roig and Xavier Boix, participate in RoboCup 2013. Polykum, an ETH newspaper, printed an article about their work. 
Read more


Best Paper Award at ECCV 2012 
The Best Paper Award for the ECCV 2012 conference was awarded to Daniel Kuettel, Dr. Matthieu Guillaumin and Dr. Vittorio Ferrari for their paper "Segmentation Propagation in ImageNet". 
Read more


Cars2 and the Computer Vision Lab 
Spin-off Procedural's CityEngine software was used in Cars2 to generate a massive number of 3D building models. 
Read more


CVL spin-off company Procedural acquired by Esri 
Esri has acquired (Procedural), the CVL spin-off founded by former PhD student Pascal Mueller and one of the world's leading software companies for creating 3D urban environments. 
Read more: from ersi.com (in English) and from ETH Life (in German).


CVL in the Swiss radio 
Prof. Luc Van Gool and the Computer Vision Laboratory were protagonists in the DRS 2 program Wissenschaft on April 30th. Hear the program (in German).


Prof. Luc Van Gool interviewed by the Swiss TV 
The program Einstein featured an interview with Prof. Luc Van Gool on March 3rd. Watch the video (in German).