Workshop on Large Data, Computer Vision and Graphics

Workshop on Large Data, Computer Vision and Graphics  

Conference Name: Workshop on Large Data, Computer Vision and Graphics

Time: June26-27, 2017

Host: School of Mathematical Science, Dalian University of Technology

Conference theme: With the popularity of various 2D and 3D graphics acquisition devices, massive multimodal data brings new opportunities and challenges to computer vision and computer graphics. The conference will invite relevant scholars to introduce the application and challenges of large data in computer vision and graphics, report the latest research results, exchange views on common issues of interest and explore possible future cooperative research. At the same time, give chance to the relevant young scholars and doctoral students to understand the relevant issues of modern frontier and the latest development opportunities.

Reporters and related information：

Reporter: Gisela Klette

Title: Skeletons in Digital Image Analysis

Abstract: Skeletons have a long history in mathematics and digital image analysis. They play an important role in many applications, often with the aim to simplify the study of digital objects by reducing the complexity of shapes prior to dedicated image analysis tasks. For 2D or 3D shapes, various skeleton types have been defined and many algorithms exist for their computation. Each method is using specific basic concepts and assumptions, and comes typically with some limitations. This talk provides a comparative discussion of different approaches for computing skeletons, their basic concepts, and a particular application in biomedical imaging showing the generation of curve skeletons in 3D digital space.

The brief introduction to the reporter: Gisela Klette is an honorary member of the Centre for Robotics & Vision at Auckland University of Technology (AUT); a Senior Lecturer at this university in Auckland, New Zealand. She authored a book (Klette, G.: Skeletal Curves in Digital Image Analysis. VDM, Saarbrücken, 2010), multiple journal and conference publications (e.g. Klette, G.: Recursive Computation of Minimum-Length Polygons. Computer Vision and Image Understanding, Volume 117, Issue 4, pages 386 - 392, 2012), and presented invited lectures at institutions in several countries. She also wrote two book chapters, one on the Euclidean distance transform, and one on relative convex hulls for 2-dimensional shape analysis.

Reporter: Baojun Li

Title: Issues on model understanding of automobile (Concept and Structure) model

Abstract: With the rapid development of computer aided design and its extensive application in industry, the digital model of complex products has been a big data trend. Based on the whole process of automobile design and manufacturing, this paper discusses the comprehension of data driven vehicle (concept and structure) model, and related research and software development of fast (automatic) modeling, simulation optimization and 3D printing from the perspective of integrated and intelligent design .

The brief introduction to the reporter: Doctor of Science, Master tutor. He is a member of International Society for Computational Mechanics (IACM), China Computer Society (CCF), the American Society of Computer Science (ACM) and the International Association of Electrical and Electronics Engineers (IEEE). Also is a director of Dalian Institute of Materials Manufacturing, executive vice director of Virtual Engineering of Virtual Engineering of Dalian University of Technology Center (National) and the vice director Dalian University of Technology Institute of automotive digital. He majors in computer aided design and graphics, CAD / CAE integration, rapid modeling based on variants and so on.

Reporter: Reinhard Klette

Title: Towards Autonomous Driving: Vision-based Driver Assistant Systems

Abstract: Prof. Reinhard Klette (Auckland University of Technology, Fellow of the Royal Society of New Zealand) made significant contributions to digital geometry and computer vision.  He is the director of the Centre for Robotics & Vision (CeRV). He has become internationally renouned for his work in vision-based driver assistance since 2006, with important contributions on performance evaluation and improvements of correspondence algorithms (for stereo matching and optical flow) on real-world video data, supporting, for example, 3D scene reconstruction from a mobile platform. Intelligence between 2001 and 2008. He is a steering committee member of the biennial conferences on Computer Analysis of Images and Patterns and of the Pacific-Rim Symposium on Image and Video Technology.

The brief introduction to the reporter: The talk will discuss computer vision challenges in the context of vision-based driver assistant systems, one of the most difficult, but also most dynamically developing areas of current 3D image Various computer-vision modules have reached the state of being robust under various conditions. The talk informs about current work in the .enpeda.. project at Auckland University of Technology (AUT), Centre for Robotics & Vision (CeRV), directed on adaptive and intelligent solutions for vision-based driver assistance or autonomous driving. The talk also reports about the use of an extensive testbed for autonomous vehicles currently established North of Auckland.

Reporter: Jinshan Pan

Title: A simple and effective blind image deblurring method

The brief introduction to the reporter: In this talk, I first introduce a simple and effective blind image deblurring method based on the dark channel prior. Our work is inspired by the interesting observation that the dark channel of blurred images is less sparse. However, sparsity of the dark channel introduces a non-convex non-linear optimization problem. We introduce a linear approximation of the min operator to compute the dark channel. Our look-up-table-based method converges fast in practice and can be directly extended to non-uniform deblurring. Extensive experiments show that our method achieves state-of-the-art results on deblurring natural images and compares favorably methods that are well-engineered for specific scenarios. Finally, I will briefly introduce our recent work on deep learning based low-level vision tasks.

Reporter:Risheng Liu

Title: Learning to Design Intelligent PDEs for Computer Vision

Abstract: Partial differential equations (PDEs) have been used to formulate image processing for several decades.  We propose a new PDE framework, named learning to diffuse (LTD), to adaptively design the governing equation and the boundary condition of a diffusion PDE system for various vision tasks on different types of visual data. To our best knowledge, the problems considered in this paper (i.e., saliency detection and object tracking) have never been addressed by PDE models before. Experimental results on various challenging benchmark databases show the superiority of LTD against existing state-of-the-art methods for all the tested visual analysis tasks.

The brief introduction to the reporter: Prof. Liu is associate Professor of School of International Information and Software, Dalian University of Technology. In recent years, 58 papers have been published in the fields of computer science and multimedia technology (T-PAMI, T-NNLS, Machine Learning, Pattern Recognition, Neural Networks, etc.) and conferences (CVPR, NIPS, AAAI, ECCV, ACCV, etc.) . He Won the Ministry of Education Natural Science Award (ranked No. 3) and "Youth Science and Technology Star" of Dalian, and he was selected to the National Ministry of Human Resources and Social Security "Xiangjiang Scholars" program and Dalian University of Technology "Xinghai Scholars" program.

Reporter: Hui Wang 

Title: Deep learning on 3D geometric shapes

The brief introduction to the reporter: Prof. Wang is a associate professor of School of Information Science and Technology, Shijiazhuang University of Railways. He was graduated from Dalian University of Technology with a degree in Computational Mathematics in 2013. His research interests include computer graphics and image processing, has published more than 10 papers. And he is selected to Railway Institutes of Shijiazhuang"outstanding science and technology youth".

Reporter: Bo Li

Title:Example-based Image Colorization using Locality Consistent Sparse Representation

Abstract:Image colorization aims to produce a natural looking color image from a given grayscale image, which remains a challenging problem. In this paper, we propose a novel example- based image colorization method exploiting a new locality consistent sparse representation. Given a single reference color image, our method automatically colorizes the target grayscale image by sparse pursuit. To the best of our knowledge, this is the first work on sparse pursuit image colorization from single reference images. Experimental results demonstrate that our colorization method outperforms state-of- the-art methods, both visually and quantitatively using a user study

The brief introduction to the reporter: Prof. Li is a associate professor of School of Mathematics and Information Science, Nanchang University of Aeronautics. His main research directions are the image processing and pattern recognition. In recent years , he has published more than 10 papers on the T-MM, Computer & Graphics and Neurocomputin.