Springer link:


Day 1 Part 1.         Part 2.         Part 3.

Day 2 Part 1.         Part 2.         Part 3.

Day 3.

Presentation guidelines

※Oral Presentation: The organizing committee will provide a computer and a video projector for display in the meeting room. The presenting author is recommended to arrive at the room 10 minutes before the start of session, and copy the presentation material (PPT or PDF) into the computer provided by the organizing committee. Each oral paper has 20 minutes, including 15 minutes for presentation and 5 minutes for questions and discussions.
※Poster Presentation: The poster boards will be attached with the paper ID. The presenting author should have the poster attached on the board of assigned number. If the authors would like to show demonstrations, they should use their own laptop with battery. The organizer will not provide desks and electricity for poster presentations. The author is recommended to prepare a poster of A0 size portrait (119cm high and 84cm wide) or smaller in portrait orientation.

August 17 (Friday)
9:00-10:00 Invited talk 1, Edwin Hancock
10:00-10:30 Coffee Break
10:30-12:10 Oral Session 1: Dissimilarity Representations (4), Gaussian Processes (1)
Paper ID: 14, 31, 36, 72, 47
12:10-14:00 Reception
14:00-16:00 Oral Session 2: Classification and Clustering(2),Deep Learning(2), Semi and Fully Supervised Learning(2)
Paper ID: 35, 61, 34, 48, 19, 55
16:00-16:30 Coffee Break
16:30-17:30 Oral Session 3: Multimedia Analysis and Understanding(3)
Paper ID: 8, 23, 49
August 18 (Saturday)
9:00-10:00 Invited talk 2, Josef Kittler
10:00-10:30 Coffee Break
10:30-12:10 Oral Session 4: Structural Matching(5)
Paper ID: 3, 10, 17, 21, 67
14:00-15:00 Invited talk 3, Xilin Chen
15:00-15:30 Coffee Break
15:30-16:30 Poster Session (19)
Paper ID: 5, 6, 11, 12, 13, 15, 22, 29, 30, 40, 41, 45, 46, 50, 54, 64, 70, 71, 74
TC1+TC2 meetings
16:30-17:30 Oral Session 5: Spatio-temporal Pattern Recognition(3)
Paper ID: 51,57,59
18:00-20:30 Banquet
August 19 (Sunday)
9:00-10:20 Oral Session 6: Graph-Theoretic Methods(4)
Paper ID: 24, 27, 37, 44
10:20-10:50 Coffee Break
10:50-12:10 Oral Session 7: Graph-Theoretic Methods(4)
Paper ID: 53, 56, 58, 63
14:00-18:00 Excursion to the Summer Palace
Invited Talk

Invited Speaker: Edwin Hancock
Institution; Department of Computer Science, University of York, UK
Time: Friday, August 17, 09:00-10:00
Room: Yixiang Hall, 2F

I began my research career in 1977 working in the field of High Energy Physics. After about eight years, I switched field to work on computer vision and pattern recognition. Soon after I made the switch, I first met Pierre Devijver who introduced me to Markov Fields. Along with Josef Kittler he proved to be one of the formative influences in my newly chosen field of research. Recently though, I find myself publishing in the physics literature again as my interests have diversified to include network science - a natural step from statistical and structural pattern recognition. This lecture seems like a good opportunity to reflect on my intellectual journey over the past 40 years. So I have borrowed J.R. Tolkein's subtitle of the Hobbit. No dragons or magic swords, but a short persons view of a massively exciting and now very topical field, where ideas from physics are rendering problems in big data tractable - allowing machine learning and pattern recognition to achieve traction in important domains such as finance and medicine.

Invited Speaker: Josef Kittler
Institution: Centre for Vision, Speech and Signal Processing, University of Surrey, UK
Time: Saturday, August 18, 09:00-10:00
Room: Yixiang Hall, 2F

Given an image, or a video, it is pertinent to ask what the semantic information content of the data is. Clearly, the data conveys information about the scene objects, but which aspect of the information is meaningful, and how can it be segregated from nuisance factors such as lighting, object pose, noise, blur and the effect of low resolution This raises two fundamental questions: i) How is semantic information defined? ii) How should semantic information be measured? Regarding the first question, we start the discussion from the perspective of philosophical theories of semantic information and consider how they map on the task of semantic information extraction from perceptual data for the benefit of an observer. The second question explores the relevance of Shannon information theory as a framework for semantic information extraction, considering that it has been developed for information transmission, rather than decision-making. The discussion will include examples, such as the problem of classifier incongruence detection, where existing information measures do not meet the requirements of the problem. We shall demonstrate that alternative information measures may be needed for specific decision making tasks and illustrate how they can be developed.

Title: Towards scene understanding
Invited Speaker: Xilin Chen
Institution: Institute of Computing Technology, Chinese Academy of Sciences (CAS), Beijing.
Time: Saturday, August 18, 14:00-15:00
Room: Yixiang Hall, 2F

Although object recognition has made a big progress in various tasks, such as face recognition, car recognition, etc. it’s still a far from the goal of automatic understanding scene. To understand scene means the machine should not only know what about an object, but also why / how / … / about an object and also their relationship in real world. I will discuss some of our recent work on this purpose. As an object exists in an ecosystem related to its property from many aspects, we model object recognition as a similarity measurement in high dimensional semantic space. This representation provides a natural hierarchical way to describe object, and it’s easy to support further tasks, such as zero shot object understanding, image caption and visual QA. The representation also provides a chance to share (latent) attributes in a common semantic space. Therefore, it can be naturally used to inference unseen classes, understanding relationship among objects, etc.

Accepted Papers

Paper ID 3 Carlos Francisco Moreno-García and Francesc Serratosa. Modelling the Generalised Median Correspondence through an Edit Distance.
Paper ID 5 Abdessalem Bouzaieni and Salvatore Tabbone. Image annotation using a semantic hierarchy.
Paper ID 6 Francesc Serratosa and Pep Santacruz. Learning the Graph Edit Distance edit costs based on an embedded model.
Paper ID 8 Jian-Xun Mi, Quanwei Zhu and Zhiheng Luo. Matrix Regression-based Classification for Face Recognition.
Paper ID 10 David Blumenthal, Sébastien Bougleux, Johann Gamper and Luc Brun. Ring Based Approximation of Graph Edit Distance.
Paper ID 11 Lichi Zhang, Han Zhang, Islem Rekik, Yaozong Gao, Qian Wang and Dinggang Shen. Malignant Brain Tumor Classification using the Random Forest Method.
Paper ID 12 Frédéric Rayar and Seiichi Uchida. On Fast Sample Preselection for Speeding up Convolutional Neural Network Training.
Paper ID 13 Mostafa Darwiche, Romain Raveaux, Donatello Conte and Vincent T'Kindt. Graph Edit Distance in the exact context.
Paper ID 14 Antonelli Mensi, Manuele Bicego, Pietro Lovato, Marco Loog and David Tax. Protein Remote Homology Detection using Dissimilarity-based Multiple Instance Learning.
Paper ID 15 Vaclav Remes and Michal Haindl. Rotationally Invariant Bark Recognition.
Paper ID 17 Vincenzo Carletti, Pasquale Foggia, Antonio Greco, Alessia Saggese and Mario Vento. The VF3-Light Subgraph Isomorphism Algorithm: when doing less is more effective.
Paper ID 19 Frank-Michael Schleif, Christoph Raab and Peter Tino. Sparsification of Indefinite Learning Models.
Paper ID 21 Xavier Cortés, Donatello Conte, Hubert Cardot and Francesc Serratosa. A Deep Neural Network Architecture to Estimate Node Assignment Costs for the Graph Edit Distance.
Paper ID 22 Xavier Cortés, Donatello Conte and Hubert Cardot. Bags of Graphs for Human Action Recognition.
Paper ID 23 Richard Wilson and Edwin Hancock. Plenoptic Imaging for Seeing Through Turbulence.
Paper ID 24 Richard Wilson and Enes Algul. Categorization of RNA Molecules using Graph Methods.
Paper ID 27 Jianjia Wang, Richard Wilson and Edwin Hancock. Quantum Edge Entropy for Alzheimer's Disease Analysis.
Paper ID 29 Cai Meng, Qi Wang, Shaoya Guan and Yi Xie. Weighted Local Mutual Information for 2D-3D Registration in Vascular Interventions.
Paper ID 30 Ryosuke Odate, Hiroshi Shinjo, Yasufumi Suzuki and Masahiro Motobayashi. Semi-supervised Clustering Framework Based on Active Learning for Real Data.
Paper ID 31 Xin Zong. Local Binary Patterns based on Subspace Representation of Image Patch for Face Recognition.
Paper ID 34 Xinran Wang, Peng Ren, Leijian Yu, Lirong Han and Xiaogang Deng. UAV First View Landmark Localization via Deep Reinforcement Learning.
Paper ID 35 Hongliu Cao, Simon Bernard, Laurent Heutte and Robert Sabourin. Dynamic voting in multi-view learning for radiomics applications.
Paper ID 36 Frédéric Rayar and Seiichi Uchida. An image-based representation for graph classification.
Paper ID 37 Nicolas Boria, Sebastien Bougleux and Luc Brun. Approximating GED using a Stochastic Generator and Multistart IPFP.
Paper ID 40 Ventzeslav Valev, Nicola Yanev, Adam Krzyzak and Karima Ben Suliman. Supervised Classification Using Feature Space Partitioning.
Paper ID 41 Lei Zhou, Shuai Wang, Xiao Bai, Jun Zhou and Edwin Hancock. Iterative Deep Subspace Clustering.
Paper ID 44 Paul Maergner, Vinaychandran Pondenkandath, Michele Alberti, Marcus Liwicki, Kaspar Riesen, Rolf Ingold and Andreas Fischer. Offline Signature Verification by Combining Graph Edit Distance and Triplet Networks.
Paper ID 45 Yun Liu, Cheng Yan and Xiao Bai. Cross-model Retrieval with Reconstruct Hashing.
Paper ID 46 Antonio Robles-Kelly, Ran Wei and Jose Alvarez. Context Free Band Reduction Using a Convolutional Neural Network.
Paper ID 47 Lorenzo Bottarelli and Marco Loog. Gradient Descent for Gaussian Processes Variance Reduction.
Paper ID 48 évariste Daller, Sebastien Bougleux, Luc Brun and Olivier Lézoray. Local Patterns and Supergraph for Chemical Graph Classification with Convolutional Networks.
Paper ID 49 Xueni Zhang, Lei Zhou, Xiao Bai and Edwin Hancock. Deep Supervised Hashing with Information Loss.
Paper ID 50 Peng Sun, Wenzhong Tang and Xiao Bai. Learning Deep Embeddings via Margin-based Discriminate Loss.
Paper ID 51 Ibrahim Caglar and Edwin Hancock. Graph Time Series Analysis using Transfer Entropy.
Paper ID 53 Giulia Sandi, Sebastiano Vascon and Marcello Pelillo. On Association Graph Techniques for Hypergraph Matching.
Paper ID 54 Zhihong Zhang, Zhuobin Xu, Zhiling Ye, Yiqun Hu, Lu Bai and Lixin Cui. Single Image Super Resolution via Neighbor Reconstruction.
Paper ID 55 Xiang Wang, Chen Wang, Xiao Bai, Yun Liu and Jun Zhou. Deep Homography Estimation with Pairwise Invertibility Constraint.
Paper ID 56 Meihong Wu, Yangbin Zeng, Zhihong Zhang, Haiyun Hong, Zhuobin Xu, Lu Bai, Lixin Cui and Edwin Hancock. Directed Network Analysis using Transfer Entropy Component Analysis.
Paper ID 57 Yuhang Jiao, Lixin Cui, Lu Bai and Yue Wang. Analyzing Time Series from Chinese Financial Market Using A Linear-Time Graph Kernel.
Paper ID 58 Linxin Cui, Lu Bai, Luca Rossi, Zhihong Zhang, Lixiang Xu and Edwin Hancock. A Mixed Entropy Local-Global Reproducing Kernel for Attributed Graphs.
Paper ID 59 Lixin Cui, Lu Bai, Luca Rossi, Zhihong Zhang, Yuhang Jiao and Edwin Hancock. A Preliminary Survey of Analyzing Dynamic Time-varying Financial Networks Using Graph Kernels.
Paper ID 61 Guangliang Chen. A scalable spectral clustering algorithm based on landmark-embedding and cosine similarity.
Paper ID 63 Manuel Curado Navarro, Francisco Escolano, Miguel Angel Lozano and Edwin Hancock. Dirichlet Densifiers: Beyond Constraining the Spectral Gap.
Paper ID 64 Shri Prakash Dwivedi and Ravi Shankar Singh. Error-Tolerant Geometric Graph Similarity.
Paper ID 67 Rafael de O. Werneck, Romain Raveaux, Salvatore Tabbone and Ricardo Da S. Torres. Learning Cost Functions for Graph Matching.
Paper ID 70 Jianming Liu, Yongsheng Gao and Yue Li. Few-Example Affine Invariant Ear Detection in the Wild.
Paper ID 71 Suhad Lateef Al-Khafaji, Jun Zhou and Alan Wee-Chung Liew. An Efficient Method for Boundary Detection from Hyperspectral Imagery.
Paper ID 72 Ziwei Xiong, Nan Zhao, Chenglong Li and Jin Tang. Visual Tracking via Patch-based Absorbing Markov Chain.
Paper ID 74 Aysylu Gabdulkhakova, Maximilian Langer, Bernhard W. Langer and Walter G. Kropatsch. Line Voronoi Diagram using Elliptical Distances.