Previous | 
Home
Section: Bilateral Contracts and Grants with Industry
Grants with Industry
CIFRE contract with Orange on Generalized lifting for video compression
Participants : Christine Guillemot, Bihong Huang.
Research axis : § 6.3.4 .
This contract with Orange labs. (starting in April. 2012) concerns the PhD of Bihong Huang and aims at investigating various forms of generalized lifting as efficient lossless and lossy coding operators in a video encoder. The generalized lifting is an extension of the lifting scheme of classical wavelet transforms which permits the creation of nonlinear and signal probability density function (pdf) dependent and adaptive transforms. This study is also carried out in collaboration with UPC (Ph. Salembier) in Barcelona.
CIFRE contract with Orange on 3D quality assessment
Participants : Darya Khaustova, Olivier Le Meur.
Research axis : § 6.1.1 .
This contract with Orange labs. (starting in Dec. 2011) concerns the PhD of Darya Khaustova and aims at developping a video quality metric for 3D content. The usage of 3D video is expected to increase in the next years. In order to ensure a good QoE (Quality of Experience), the 3D video quality must be monitored and accuratly measured. The goal of this thesis is to study objective measures suitable for estimating 3D video quality. A comparison with ground truth as well as with the state-of-the-art 2D metrics should be carried out. To be as effective as possible, the feature of the human visual system should be taken into account.
CIFRE contract with Technicolor on sparse modelling of spatio-temporal scenes
Participants : Martin Alain, Safa Cherigui, Christine Guillemot.
Title : Spatio-temporal analysis and characterization of video scenes
Research axis : § 6.1.3 .
Period : Ph.D Of S. Cherigui (Nov.09- Oct.12); Ph.D. of M. Alain (Oct.12-Sept.15).
A first CIFRE contract has concerned the Ph.D of Safa Cherigui from Nov.2009 to Oct.2012, in collaboration with Dominique Thoreau (Technicolor). The objective was to investigate texture and video scene characterization methods and models based on sparse and data dimensionality reduction techniques, as well as on concepts of epitomes. The objective was then to use these models and methods in different image processing problems focusing in particular on video compression. Among the main achievements, one can first cite novel Intra prediction techniques using data dimensionality reduction methods with correspondance map-aided K-NN search. A novel method has also been introduced for constructing epitome representations of an image, with rate-distortion optimization criteria. The epitome coupled with inpainting is used for image compression, showing significant performance gains with respect to H.264 Intra prediction modes. The above results have led to a best student paper award at IEEE-ICASSP 2012 [21] , one nominated paper among the 11 finalists for a best student paper award at IEEE-ICIP 2012 [22] , one journal publication in the IEEE Trans. on Image Processing [13] and to 6 patents.
The first PhD thesis has focused on spatial analysis, processing, and prediction of image texture. A second CIFRE contract (PhD thesis of Martin Alain) has hence just started to push further the study by addressing issues of spatio-temporal analysis and epitome construction, with applications to temporal prediction, as well as to other video processing problems such as video inpainting.
CIFRE contract with Thomson Video Networks (TVN) on Video analysis for HEVC based video coding
Participants : Nicolas Dhollande, Christine Guillemot, Olivier Le Meur.
This contract with TVN (starting in Oct. 2012) concerns the PhD of Nicolas Dhollande and aims at performing a coding mode analysis and developing a pre-analysis software. HEVC standard is a new standard of compression including new tools such as advanced prediction modes. Compared to the previous standard H.264, HEVC's complexity is three to four times higher. The goal of this thesis is to to infer the best coding decisions (prediction modes...) in order to reduce the computational complexity of HEVC. A pre-analysis or a first coding pass. This latter would provide useful estimates such as rate distortion characteristics of video, the set of quantization parameters (often called just QPs). These estimated characteristics would help to select quantization parameters for the second pass in order to reduce the computational complexity by keeping a good quality.