EN FR
EN FR


Section: Bilateral Contracts and Grants with Industry

Bilateral Grants with Industry

Consulting contract with Enensys technologies

Participant : Aline Roumy.

  • Title : Matrix inversion for video streaming.

  • Research axis : REF AT 7.4. Distributed processing and robust communication

  • Partners : Enensys, Inria-Rennes.

  • Funding : Enensys.

  • Period : Apr. 2016 - May 2016.

This contract with Enensys technologies aimed at studying solutions for reducing the complexity of matrix inversion used for encoding data in the context of video streaming. First a bibliographical study has been carried out related to the problem of matrix inversion in a finite field, then a novel solution has been proposed together with some recommendations regarding the algorithmic implementation.

Google faculty research award

Participants : Christine Guillemot, Xiaoran Jiang, Mikael Le Pendu.

  • Title  : Light fields low rank and sparse approximation

  • Research axis : 7.3.2

  • Partners : Inria-Rennes.

  • Funding : Google.

  • Period : Oct.2015-Sept.2016.

The goal of the project was to study low-rank and sparse approximation models for light fields compression. A homography-based low-rank approximation has been developed showing significant PSNR-rate gains compared to a direct encoding of light field views with HEVC-inter coding.

CIFRE contract with Envivio/ Ericsson on LDR compatible HDR video coding

Participants : Christine Guillemot, David Gommelet, Aline Roumy.

  • Title  : LDR-compatible coding of HDR video signals.

  • Research axis : § 7.3.5.

  • Partners : Envivio.

  • Funding : Cifre Envivio.

  • Period : Oct.2014-Sept.2017.

The goal of this Cifre contract is to design solutions for LDR-compatible coding of HDR videos. This involves the study of rate-distortion optimized tone mapping operators taking into account constraints of temporal coherency to avoid the temporal flickering which results from a direct frame-by-frame application of classical tone mapping operators. The goal is also to design a coding architecture which will build upon these operators, integrating coding tools tailored to the statistics of the HDR refinement signals.

CIFRE contract with Harmonic on image analysis for HDR video compression

Participants : Maxime Rousselot, Olivier Le Meur.

  • Title : image and video analysis for HDR video compression

  • Partners : Harmonic, Univ. Rennes 1

  • Funding: Harmonic, ANRT

  • Period: April 2016-April 2019

This project (in collaboration with Rémi Cozot, FRVSense) aims to investigate two main axes. First, we want to assess whether the representation of High Dynamic Range signal has an impact on the coding efficiency. We will focus mainly on the Hybrid Log-Gamma (HLG) and Perceptual Quantizer (PQ) OETF (Opto-Electronic Transfer Function)approaches. The former defines a nonlinear transfer function which is display-independent and able to produce high quality images without compromising the director’s artistic intent. The latter approach is based on Just Noticeable Difference curve. If it turns out that this representation has an impact, the coding strategy should be adjusted with respect to the representation. In addition, specific preprocessing tools will be defined to deal with the limitations of PQ and HLG approaches.

CIFRE contract with Technicolor on image collection analysis

Participants : Dmitry Kuzovkin, Olivier Le Meur.

  • Title : Spatiotemporal retargeting and recomposition based on artistic rules

  • Partners : Technicolor, Univ. Rennes 1

  • Funding: Technicolor, ANRT

  • Period: Nov. 2015 – Nov. 2018

The goal of the project (in collaboration with Rémi Cozot, FRVSense) is to take advantage of the huge quantities of image and video data currently available - captured by both amateur and professional users - as well as the multiple copies of each scene that users often capture, to improve the aesthetic appeal of content. Additionally, given Technicolor's unique position, we propose to take advantage of insights as well as content from professional artists and colorists to learn how different content types can be enhanced.

CIFRE contract with Technicolor on light fields editing

Participants : Christine Guillemot, Matthieu Hog.

  • Title  : Light fields editing

  • Research axis : 7.1.2

  • Partners : Technicolor, Inria-Rennes.

  • Funding : Technicolor, ANRT.

  • Period : Oct.2015-Sept.2018.

Editing is quite common with classical imaging. Now, if we want light-fields cameras to be in the future as common as traditional cameras, this functionality should also be enabled with light-fields. The goal of the PhD is to develop methods for light-field editing, and the work in 2016 has focused on the design of fast semi-supervised segmentation algorithms with coherence constraints across sub-aperture images (see 7.1.2).

CIFRE contract with Technicolor on light fields compressed representation

Participants : Christine Guillemot, Fatma Hawary.

  • Title  : Light fields compressed representation

  • Partners : Technicolor, Inria-Rennes.

  • Funding : Technicolor, ANRT.

  • Period : Feb.2016-Jan.2019.

The goal of this PhD is to study reconstruction algorithms from compressed measurements based on the assumption of sparsity in the Fourier domain. The goal is to apply these algorithms to scalable compression of light fields.

CIFRE contract with Technicolor on cloud-based image compression

Participants : Jean Begaint, Christine Guillemot.

  • Title  : Cloud-based image compression

  • Research axis : 7.3.6

  • Partners : Technicolor, Inria-Rennes.

  • Funding : Technicolor, ANRT.

  • Period : Nov.2015-Oct.2018.

The goal of this Cifre contract is to develop a novel image compression scheme exploiting similarity between images in a cloud. The objective will therefore be to develop rate-distortion optimized affine or homographic estimation and compensation methods which will allow us to construct prediction schemes and learn adapted bases from most similar images retrieved by image descriptors. One issue to be addressed is the rate-distortion trade-off induced by the need for transmitting image descriptors.