- Prof. Dong Xu, School of Computer Engineering, Nanyang Technological University, Singapore
- Title: Dimensionality Reduction: From General Formulation to Data Representationi
- Date: January 5, 2015 Time: 10:50-11:50 Hrs Venue: Platinum Jubilee Auditorium
- Lecture Slides
-
Prof. Fernand Meyer, Directeur de recherche, CMM, Fontainebleau, Mines ParisTech, France
- Title: Stochastic Watershed Hierarchies
- Date: January 6, 2015 Time: 12:00-13:00 Hrs Venue: Platinum Jubilee Auditorium
Abstract:
Over the past few decades, a large family of algorithms has been designed to provide different solutions to the problem of dimensionality reduction. In this talk, I will first describe a general formulation known as graph embedding to unify them within a common framework. Furthermore, the graph embedding framework can be used as a general platform for developing new dimensionality reduction algorithms. I will then introduce a series of new dimensionality reduction algorithms developed under this framework, in which image objects are represented in their intrinsic forms and orders (e.g., an image is a second-order tensor (i.e., a matrix) and sequential data such as video sequences used in event analysis is in the form of a third-order tensor). Moreover, I will also present the applications of our dimensionality reduction algorithms for face recognition and human gait recognition.
Brief CV: Dr. Dong Xu is currently an associate professor in the School of Computer Engineering, Nanyang Technological University, Singapore. His research focuses on new theories, algorithms and systems for intelligent processing and understanding of visual data such as images and videos. One of his co-authored works on domain adaptation for video event recognition won the Best Student Paper Award in CVPR 2010. He is on the editorial boards of IEEE Transactions on Pattern Analysis and Machine Intelligence, IEEE Transactions on Neural Networks and Learning Systems, and Machine Vision and Applications (Springer).
Abstract: We present a segmentation strategy which first constructs a hierarchy, i.e. a series of nested partitions. A coarser
partition is obtained by merging adjacent regions in a finer partition. The strength of a contour is then measured by the level of the hierarchy for which its two adjacent regions merge. Various strategies are presented for constructing hierarchies which highlight specific features of the image. The last part shows
how the hierarchies lead to a final segmentation.
Brief CV: Fernand Meyer got an engineer degree from the Mines-ParisTech in 1975. He works since 1975 at the Centre de Morphologie Mathematique (CMM) of Mines-ParisTech (member of PSL Research University). His first research area was ''Early and Automatic Detection of Cervical Cancer on Cytological Smears'', subject of his PhD thesis, obtained in 1979. He participated actively to the development of mathematical morphology: reconstruction openings, top-hat transform, the morphological segmentation paradigm based on the watershed transform and markers, the theory of digital skeleton, the introduction of hierarchical queues for high speed watershed computations, morphological interpolations, the theory of levelings, hierarchical segmentation. His current subject of interest is the extension of mathematical morphology to node and/or edge weighted graphs.