Corpus ID: 18613255

Group integration techniques in pattern analysis - a kernel view

  title={Group integration techniques in pattern analysis - a kernel view},
  author={Marco Reisert},
  booktitle={Ausgezeichnete Informatikdissertationen},
  • M. Reisert
  • Published in
  • Computer Science
Pattern analysis deals with the problem of characterizing a nd analyzing relations in data in an automated way. A quite important issue during the design p rocess of such algorithms is the incorporation of prior knownledge; knowledge that is related to all information about the problem available in addition to the training data . This thesis is about a certain kind of prior knowledge: we assume to know that the data does n t change its meaning under certain transformations, that is, the pattern… Expand
Discriminative Learning of Similarity and Group Equivariant Representations
A SO(3)-equivariant neural network architecture for spherical data, that operates entirely in Fourier space, while also providing a formalism for the design of fully Fourier neural networks that are equivariant to the action of any continuous compact group. Expand
Unitary-Group Invariant Kernels and Features from Transformed Unlabeled Data
This paper studies kernels that are invariant to the unitary group while having theoretical guarantees in addressing practical issues such as unavailability of transformed versions of labelled data and not observing all transformations. Expand
Max-Margin Invariant Features from Transformed Unlabelled Data
This paper designs an framework for face recognition and demonstrates the efficacy of the approach on a large scale semi-synthetic dataset with 153,000 images and a new challenging protocol on Labelled Faces in the Wild (LFW) while out-performing strong baselines. Expand
Training or Architecture? How to Incorporate Invariance in Neural Networks
This work proposes a method for provably invariant network architectures with respect to group actions by choosing one element from a (possibly continuous) orbit based on a fixed criterion and intends to ’undo’ any possible transformation before feeding the data into the actual network. Expand
DRAFT VERSION of PART II : Invertible Orientation Scores
Inspired by the early visual system of many mammalians we consider the construction of-and reconstruction froman orientation score Uf : R × Sd−1 → C as a local orientation representation of an image,Expand
Universal approximations of invariant maps by neural networks
The "charge--conserving convnet" -- a convnet-like computational model based on the decomposition of the feature space into isotypic representations of SO(2) is introduced and it is proved to be a universal approximator for continuous SE(2)--equivariant signal transformations. Expand
Design and Processing of Invertible Orientation Scores of 3D Images
3D versions of the 2D cake wavelets, which are complex wavelets that can simultaneously detect oriented structures and oriented edges are introduced, and a new tubularity measure is developed and applied to both artificial and real medical data. Expand
From voxels to models: towards quantification in 3-D confocal microscopy
Current microscopic imaging techniques allow to record biological specimen with sub-micrometer resolution in 3-D (over time). They open completely new ways to observe and understand signaling inExpand
RENNSH: A Novel \alpha-Helix Identification Approach for Intermediate Resolution Electron Density Maps
A novel refined classification framework is proposed, which treats alpha-helix identification as a machine learning problem by representing each voxel in the density map with its Spherical Harmonic Descriptors. Expand
Increasing the Dimension of Creativity in Rotation Invariant Feature Design Using 3D Tensorial Harmonics
This work proposes a novel approach extending the spherical harmonic representation to tensors of higher order in a very efficient manner and utilises the so called tensorial harmonics to overcome the restrictions to scalar fields. Expand


Invariant kernel functions for pattern analysis and machine learning
This work presents two generic approaches for constructing invariant kernels and proposes a more distinguishing treatment in particular in the active field of kernel methods for machine learning and pattern analysis, to enable a smooth interpolation between invariant and non-invariant pattern analysis. Expand
Support vector learning
This book provides a comprehensive analysis of what can be done using Support vector Machines, achieving record results in real-life pattern recognition problems, and proposes a new form of nonlinear Principal Component Analysis using Support Vector kernel techniques, which it is considered as the most natural and elegant way for generalization of classical Principal Component analysis. Expand
Kernel Methods for Pattern Analysis
This book provides an easy introduction for students and researchers to the growing field of kernel-based pattern analysis, demonstrating with examples how to handcraft an algorithm or a kernel for a new specific application, and covering all the necessary conceptual and mathematical tools to do so. Expand
Estimating the Support of a High-Dimensional Distribution
The algorithm is a natural extension of the support vector algorithm to the case of unlabeled data by carrying out sequential optimization over pairs of input patterns and providing a theoretical analysis of the statistical performance of the algorithm. Expand
Feature Detection with Automatic Scale Selection
  • T. Lindeberg
  • Computer Science
  • International Journal of Computer Vision
  • 2004
It is shown how the proposed methodology applies to the problems of blob detection, junction detection, edge detection, ridge detection and local frequency estimation and how it can be used as a major mechanism in algorithms for automatic scale selection, which adapt the local scales of processing to the local image structure. Expand
Tangent distance kernels for support vector machines
  • B. Haasdonk, Daniel Keysers
  • Mathematics, Computer Science
  • Object recognition supported by user interaction for service robots
  • 2002
This work introduces a new class of kernels for support vector machines which incorporate tangent distance and therefore are applicable in cases where such transformation invariances are known. Expand
Statistical Pattern Recognition
  • J. Davis
  • Computer Science
  • Technometrics
  • 2003
This chapter introduces the subject of statistical pattern recognition (SPR). It starts by considering how features are defined and emphasizes that the nearest neighbor algorithm achieves error ratesExpand
Adjustable invariant features by partial Haar-integration
A method for Haar-integration is generalized to make it applicable to more general transformation sets, namely subsets of transformation groups, and increased separability by these features and considerably improved recognition performance on a character recognition task is demonstrated. Expand
A group-theoretic approach to the triple correlation
  • R. Kakarala
  • Mathematics
  • [1993 Proceedings] IEEE Signal Processing Workshop on Higher-Order Statistics
  • 1993
The triple correlation is a useful tool for averaging multiple observations of a signal in noise, in particular when the signal is translating by unknown amounts in between observations. What makesExpand
Continuous Group Averaging and Pattern Classification Problems
  • J. Dunn
  • Mathematics, Computer Science
  • SIAM J. Comput.
  • 1973
The Fourier transform of a plane figure is averaged over the one-parameter continuous group of dilatations to obtain a pair of interesting scale invariant transforms which tend to pick out corners and flat spots in the figure’s boundary. Expand