19 September

KEYNOTE SPEAKER: Material Appearance Issues in Cultural Heritage (09:00—09:45)

Holly Rushmeier

Abstract. A great deal of effort by researchers in computer graphics and vision has gone into developing methods to capture the shape of historic sites and objects for documentation and study. However, geometry is not enough to thoroughly document, or to study many issues. This has driven efforts to develop material appearance capture methods that can work under adverse conditions in the field. Capturing appearance is important for studying issues such as the composition and preservation status of objects. It is also important for performing experiments in visibility in past environments to understand how sites and objects were used. This talk will survey work both in capturing material scattering properties from heritage sites and in modeling illumination. Opportunities for further work will be presented

COFFEE BREAK (09:45—10:00)

Poster Session (10:00—12:00)
Chair: Jean Baptiste Thomas
Novel approach to uniformization of a color space via generic deep learning-based transformation
  • Oleksii Sidorov; COSI program, University of Granada, Spain
Deep learning for dehazing: Comparison and analysis
  • Leonel Cuevas Valeriano; The Norwegian Colour and Visual Computing Laboratory, NTNU Norway
  • Jean-Baptiste Thomas; The Norwegian Colour and Visual Computing Laboratory, NTNU Norway
  • Alexandre Benoit; LISTIC, Univ. Savoie Mont Blanc, France
Can image quality enhancement methods improve the performance of biometric systems for degraded face images?
  • Xinwei Liu; Normandie Univ, UNICAEN, ENSICAEN, CNRS, GREYC, Caen, France, Norwegian University of Science and Technology, Gjøvik, Norway
  • Marius Pedersen; Norwegian University of Science and Technology, Gjøvik, Norway
  • Christophe Charrier; Normandie Univ, UNICAEN, ENSICAEN, CNRS, France
  • Patrick Bours; Norwegian University of Science and Technology, Gjøvik, Norway
Hand-crafted vs deep features: A quantitative study of pedestrian appearance model
  • Mohib Ullah; Department of Computer Science (IDI), NTNU, Norway
  • Mohammed Ahmed Kedir; Department of Computer Science (IDI), NTNU, Norway
  • Faouzi Alaya Cheikh; Department of Computer Science (IDI), NTNU, Norway
Melon crack identification and classification using k-means clustering for quality inspection
  • Uldanay Bairam, Norwegian University of Science and Technology, Gjøvik, Norway
  • Philip John Green, Norwegian University of Science and Technology, Gjøvik, Norway
Statistics of hyperspectral data/image analysis: Entropy
  • Yu-Jung Chen; XLIM, JRU CNRS 7252, University of Poitiers, France
  • Noël Richard; XLIM, JRU CNRS 7252, University of Poitiers, France
  • Hilda Deborah; Norwegian University of Science and Technology, Norway
  • Aurélie Tournié, CRC, NMNH, Paris, France
  • Anne Michelin; CRC, NMNH, Paris, France
  • Christine Andraud; CRC, NMNH, Paris, France
A corrected single-constant Kubelka-Munk model for color prediction of pre-colored fiber blends
  • Chun-ao Wei; School of Printing and Packaging Wuhan University, China
  • Xiaoxia Wan; School of Printing and Packaging Wuhan University, China
  • Hang Luo; School of Printing and Packaging Wuhan University, China
A key frame based video summarization using color features
  • Muhammad Asim; Qatar University, Doha, Qatar
  • Noor Almaadeed; Qatar University, Doha, Qatar
  • Somaya Al-maadeed; Qatar University, Doha, Qatar
  • Ahmed Bouridane; Department of Computer and Information Sciences, Northumbria University, UK
  • Azeddine Beghdadi; L2TI, Université Paris 13, France
Determining the sequence of intersecting lines in document forgery by colormetric evaluation
  • Atefeh Tajik Esmaeili; Department of Color Imaging, Institute for Color Science and Technology, Tehran, Iran
  • Mahdi Safi; Department of Color Imaging, Institute for Color Science and Technology, Tehran, Iran
  • Maryam Ataeefard; Department of Printing Science and Technology, Institute for Color Science and Technology, Tehran, Iran
A pilot study on Iranian skin color
  • Majid Ansari-Asl; Color Imaging Laboratory, University of Granada, Spain
  • Keivan Ansari; Department of Color Imaging & Color Image Processing, Iranian Institute for Color Science and Technology, Iran
Spatially dependent white balance for fill flash photography
  • Enrique Gurdiel; The Norwegian Colour and Visual Computing Laboratory, NTNU Norway
  • Jon Yngve Hardeberg; The Norwegian Colour and Visual Computing Laboratory, NTNU Norway
CNN feature similarity: Paintings are more self-similar at all levels
  • Seyed Ali Amirshahi; Norwegian Colour and Visual Computing Laboratory, NTNU Norway. L2TI, University Paris 13, France
  • Stella X. Yu; UC Berkeley/ ICSI, Berkeley, CA., USA
A deep learning-based human activity recognition in darkness
  • Md. Zia Uddin; Dept. of Informatics, University of Oslo, Norway
  • Jim Torresen; Dept. of Informatics, University of Oslo, Norway

LUNCH BREAK (12:00—13:00)

KEYNOTE SPEAKER: Online Machine Learning for Robot Vision (13:00—13:45)

Michael Felsberg

Abstract. With the advent of deep learning, machine learning became the dominating paradigm to address most computer vision problems. However, training deep networks usually requires enormous amounts of data and computational effort in the order of days or weeks. In robot vision applications, data is typically scarce, and systems might be exposed to completely new situations, which require instantaneous training from scratch with limited data. Standard approaches for deep learning will obviously not be applicable in these situations and thus our ongoing research focusses on combining real-time machine learning methods with suitable ingredients from deep learning. In particular off-line trained features are useful in many situations, but the fusion of multi-modal and multi-scale features is a challenging problem. Considering, e.g., colour and shape features as independent cues is often not the most powerful approach, but modelling the full combined space is not feasible either. In our methods we aim at combining advantages of both approaches. During the presentation, state-of-the-art solutions to various robot vision problems such as point-cloud registration, depth completion, object- and action detection, regression of robot control, and object tracking will be explained. Particularly our approaches to the latter problem have recently gained substantial attention in the field and our algorithms have achieved top results on several international benchmarks

COFFEE BREAK (13:45—14:00)

Vision Session (14:00—15:00)
Chair: Holly Rushmeier
CEED - A Database for image contrast enhancement evaluation
  • Azeddine Beghdadi; L2TI, Institut Galilée, Université Paris 13, Sorbonne Paris Cité, France
  • Muhammad Ali Qureshi; The Islamia University of Bahawalpur, Pakistan
  • Bilel Sdiri; L2TI, Institut Galilée, Université Paris 13, Sorbonne Paris Cité, France, NTNU Norway
  • Mohamed Deriche; King Fahd University of Petroleum and Minerals, Saudi Arabia
  • Faouzi Alaya-Cheikh; Norwegian University of Science and Technology, Gjøvik, Norway
Colour-to-Greyscale image conversion by linear anisotropic diffusion of perceptual colour metrics
  • Ivar Farup; Norwegian University of Science and Technology, Gjøvik, Norway
  • Marius Pedersen; Norwegian University of Science and Technology, Gjøvik, Norway
  • Ali Alsam; Norwegian University of Science and Technology, Trondheim, Norway
Methods for psychophysical assessment of colour difference by observers with a colour vision deficiency
  • Anne Kristin Kvitle; Norwegian University of Science and Technology, Gjøvik, Norway
  • Henrik Oddløkken; Norwegian University of Science and Technology, Gjøvik, Norway
  • Phil Green; Norwegian University of Science and Technology, Gjøvik, Norway
  • Peter Nussbaum; Norwegian University of Science and Technology, Gjøvik, Norway

COFFEE BREAK (15:00—15:20)

Color Imaging Session (15:20—16:20)
Chair: Shoji Tominaga
Comparison of mosaic patterns for spectral filter arrays
  • Davit Gigilashvili; The Norwegian Colour and Visual Computing Laboratory, NTNU, Norway
  • Jon Yngve Hardeberg; The Norwegian Colour and Visual Computing Laboratory, NTNU, Norway
  • Jean-Baptiste Thomas; The Norwegian Colour and Visual Computing Laboratory, NTNU, Norway
Dye purification: an image-processing technique for the digital restoration of chromogenic film
  • Giorgio Trumpy; Department of Film Studies, University of Zurich
  • Barbara Flueckiger; Department of Film Studies, University of Zurich
Evaluation of color correction methods for printed surfaces
  • Maliha Ashraf; Université Jean Monnet, Océ Print Logic Technologies S.A. Créteil, France
  • Luis Ricardo Sapaico; Océ Print Logic Technologies S.A. Créteil, France
INVITED PAPER: Transparent Materials - Colour and Appearance (16:20—16:45)

Patrick Callet

ColourLab Tour (16:45—17:45)

Conference Dinner (19:00—23:00)

20 September

KEYNOTE SPEAKER: From Vision Models To Cinema Applications, And Back (09:15—10:00)

Marcelo Bertalmío

Abstract. A key goal of cinema post production is to emulate perception, and cinema professionals devote a lot of effort to modify color and contrast of the raw footage so that it appears in the movie screen as it would look without extra lights or any other processing to an observer present in the scene. But despite the complexity of the user interface of professional color grading tools, the underlying processing is based on very simplistic models with parameters that are unrelated to perception, which makes the work of the colorist very challenging and time consuming or yields underwhelming results. In this talk I will describe in detail these issues, review some recent methods for contrast and color modification which are based on vision models and that can be successfully applied to these problems in cinema postproduction, and discuss how the technical expertise of cinematographers and colorists can help vision science in obtaining better models of perception

COFFEE BREAK (10:00—10:15)

Colour and Light Session (10:15—11:15)
Chair: Patrick Callet
Effects of ambient illumination on text recognition for UI development
  • Giovanni Pignoni; Department of Design, Norwegian University of Science and Technology
Evaluation of gamut mapping algorithms in different uniform colour spaces
  • Baiyue Zhao; State Key Laboratory of Modern Optical Instrument, Zhejiang University, China
  • Lihao Xu; State Key Laboratory of Modern Optical Instrument, Zhejiang University, China
  • Ming Ronnier Luo; State Key Laboratory of Modern Optical Instrument, Zhejiang University, China
  • Muhammad Safdar; Department of Computer Science COMSATS Institute of Information Technology, Pakistan
Photo-realistic style transfer for cinema shoots
  • Itziar Zabaleta; Dpt. of Information and Communication Technologies, Universitat Pompeu Fabra, Spain
  • Marcelo Bertalmıo; Dpt. of Information and Communication Technologies, Universitat Pompeu Fabra, Spain

COFFEE BREAK (11:15—11:30)

Medical Session (11:30—12:30)
Chair: Azzedine Beghdadi
Validation of stereo vision based liver surface reconstruction for image guided surgery
  • Teatini Andrea; The Intervention Centre, Oslo University Hospital, Oslo, Norway.
  • Wang Congcong; Norwegian University of Science and Technology, Gjøvik, Norway
  • Palomar Rafael; The Intervention Centre, Oslo University Hospital, Oslo; NTNU, Norway.
  • Alaya Cheikh Faouzi; Norwegian University of Science and Technology, Gjøvik, Norway
  • Beghdadi Azeddine; L2TI-Institut Galilée, Université Paris 13, Sorbonne Paris, France
  • Edwin Bjørn; The Intervention Centre, Oslo University Hospital, Oslo, Norway. Oslo University Hospital, Department of Hepato-Pancreatic-Biliary surgery, Oslo, Norway
  • Elle Ole Jakob; The Intervention Centre, Oslo University Hospital, Oslo, Norway. Department of Informatics, University of Oslo, Oslo, Norway
Stochastic correction of boundary conditions during liver surgery
  • Sergei Nikolaev; University of Strasbourg, Inria, France
  • Igor Peterlik; Inria, France
  • Stephane Cotin; Inria, France
Enhancing dermoscopy images to improve melanoma detection
  • Olga Cherepkova; The Norwegian Colour and Visual Computing Laboratory, NTNU, Norway
  • Jon Yngve Hardeberg; The Norwegian Colour and Visual Computing Laboratory, NTNU, Norway

LUNCH BREAK (12:30—13:30)

Appearance Session (13:30—14:30)
Chair: Philipp Urban
Dependence of texture classification accuracy on spectral information
  • Michele Conni; Barbieri Electronic, Italy; Department of Computer Science, NTNU, Gjøvik, Norway
  • Helene Midtfjord; Department of Computer Science, NTNU, Gjøvik, Norway
  • Peter Nussbaum; Department of Computer Science, NTNU, Gjøvik, Norway
  • Phil Green; Department of Computer Science, NTNU, Gjøvik, Norway
Measurement uncertainty for printed textiles
  • Nadile Nunes de Lima; Barbieri electronic, Italy and NTNU Gjøvik
  • Michele Conni; Barbieri electronic, Italy and NTNU Gjøvik
  • Phil Green; Norwegian University of Science and Technology, Gjøvik, Norway
  • Markus Barbieri; Barbieri electronic, Italy
Application of spectral statistics to spectral texture discrimination
  • Hilda Deborah; Norwegian University of Science and Technology, Gjøvik, Norway
  • Noël Richard; Laboratory XLIM, JRU CNRS 7252, University of Poitiers, France
  • Jon Yngve Hardeberg; Norwegian University of Science and Technology, Gjøvik, Norway

Best Student Paper Award, Closing (14:30—14:45)

downloadable version of CVCS2018 Program