Postdoctoral position on structure-preserving graph signal processing
The postdoctoral position is funded under the research project GRAPHSIP (Graph Signal Processing - P.I. : O. Lézoray, University of Caen – https://graphsip.greyc.fr) supported by the French National Research Agency and the Region Council of Normandy. Background: ----------- Processing of very large datasets with irregular structure poses significant challenges, as they can be nowa- days collected in numerous domains, from engineering sciences to social networks. Massive datasets rep- resented as graphs can be seen as a set of data samples, with one sample at each vertex in the graph. In such a scenario, the high-dimensional data associated to vertices can be viewed as graph signals. As a consequence, Graph Signal Processing (GSP) has recently emerged as a new methodology for such irreg- ular big data processing [1, 2]. However, classical signal processing methods are defined only for regular grid-like graphs (such as images ), and for irregular graphs usual signal processing methods have to be completely rethought. There is therefore a strong interest (both on theoretical and application sides) for the development of a unified theory for the processing and analysis of irregular graph signals. Description of the position: Recently, low cost sensors have brought 3D scanning into the hands of consumers and one can now easily produce 3D colored meshes or point clouds with each vertex described by its position and color (a 3D colored graph signal). However, the quality of the 3D data is not always visually good and several post-production steps are necessary to improve the final quality. Traditional image processing for editing tasks use structure-preserving smoothing filters [4, 5, 6, 7] within a hierarchical framework. Structure-preserving filters distinguish details from major image structures based on color or patch differences. Then, they decompose an image into different layers from coarse scale structures to small scale fine details, making it easier for subsequent detail manipulation (abstraction, simplification, enhancement, completion, ...). Some filters have been extended to 3D meshes but most of them merely handle vertex positions [8, 9] and cannot deal with irregular graph signals . The post-doctoral research scientist will join the image team of the GREYC laboratory located in Caen (UMR CNRS 6072 – https: //www.greyc.fr/en) to develop new methods for editing 3D colored graph signals. He will focus on the adaptation, under a variational formulation for irregular graphs, of structure-preserving signal processing methods based on data-adaptive regularization. Convolutional networks on graphs will be also be of interest [11, 12] for an efficient extraction of data-adaptive features. The developed methods will be investigated for detail manipulation of 3D selfies . Candidate profile:
The candidate must have a recent Ph.D. (within 5 years) in Computer Science or Applied Mathematics. Knowledge in the areas of graphs, image processing, computer vision or computer graphics is also very welcomed. The candidate will perform research and algorithmic development in C++ and solid programming skills are required. Excellent interpersonal skills and the ability to work well individ- ually or as a member of a project team are recommended. Good written and verbal communication skills required, the candidate has to be fluent in English both written and spoken. Working language can be English or French. Location: Caen, France in the GREYC UMR CNRS laboratory. Situated in the Normandy region of France close to the sea and about 240km west of Paris; the city still has many old quarters, a population of around 120,000; the city area has roughly 250,000 inhabitants. Duration: One year, starting in September 2017.
Possibility of French courses, participation in transport costs, possibility of restoration on site. Gross Salary: 2500 euros per month. Application: Interested applicants should submit (by email, in a single pdf file) their Curriculum Vitae, list of publications, a statement of research interests and 2 reference letters. Contact: Olivier Lézoray (firstname.lastname@example.org) – https://lezoray.users.greyc.fr. Applications will be admitted until the position is filled. References  D. I. Shuman, S. K. Narang, P. Frossard, A. Ortega, and P. Vandergheynst, “The emerging field of signal processing on graphs: Extending high-dimensional data analysis to networks and other irregular domains,” IEEE Signal Process. Mag., vol. 30, no. 3, pp. 83–98, 2013.  A. Sandryhaila and J. M. F. Moura, “Big data analysis with signal processing on graphs: Representation and processing of massive data sets with irregular structure,” IEEE Signal Processing Magazine, vol. 31, no. 5, pp. 80–90, 2014.  O. Lézoray and L. Grady, Image Processing and Analysis with Graphs: Theory and Practice, Digital Imaging and Computer Vision. CRC Press / Taylor and Francis, 2012.  Qi Zhang, Xiaoyong Shen, Li Xu, and Jiaya Jia, “Rolling guidance filter,” in European Conference on Computer Vision ECCV, 2014, pp. 815–830.  H. Cho, H. Lee, H. Kang, and S. Lee, “Bilateral texture filtering,” ACM Transactions on Graphics, vol. 33, no. 4, pp. 128:1–128:8, 2014.  E. S. L. Gastal and M. M. Oliveira, “Domain transform for edge-aware image and video processing,” ACM Transactions on Graphics, vol. 30, no. 4, pp. 69, 2011.  L. Xu, C. Lu, Y. Xu, and J. Jia, “Image smoothing via L0 gradient minimization,” ACM Transactions on Graphics, vol. 30, no. 6, pp. 174, 2011.  S. Fleishman, I. Drori, and D. Cohen-Or, “Bilateral mesh denoising,” ACM Transactions on Graphics, vol. 22, no. 3, pp. 950–953, 2003.  Michael Kolomenkin, Ilan Shimshoni, and Ayellet Tal, “Prominent field for shape processing and analysis of archaeological artifacts,” International Journal of Computer Vision, vol. 94, no. 1, pp. 89–100, 2011.  M. Hidane, O. Lézoray, and A. Elmoataz, “Graph signal decomposition for multi-scale detail manipulation,” in International Conference on Image Processing (IEEE), 2014, pp. 2041–2045.  Jonathan Masci, Davide Boscaini, Michael M. Bronstein, and Pierre Vandergheynst, “Geodesic convolutional neural networks on riemannian manifolds,” in 2015 IEEE International Conference on Computer Vision Work- shop, ICCV Workshops 2015, Santiago, Chile, December 7-13, 2015, 2015, pp. 832–840.  Michael Edwards and Xianghua Xie, “Graph based convolutional neural network,” CoRR, vol. abs/1609.08965, 2016.  O. Lézoray, “3d colored mesh graph signals multi-layer morphological enhancement,” in International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2017, vol. accepted, p. to appear.