Mode of access: World Wide Web.

Part of: Synthesis digital library of engineering and computer science.

Includes bibliographical references (pages 71-81) and index.

1. Introduction -- 1.1 Feature design -- 1.2 Channel representations: a design choice --

2. Basics of feature design -- 2.1 Statistical properties -- 2.2 Invariance and equivariance -- 2.3 Sparse representations, histograms, and signatures -- 2.4 Grid-based feature representations -- 2.5 Links to biologically inspired models --

3. Channel coding of features -- 3.1 Channel coding -- 3.2 Enhanced distribution field tracking -- 3.3 Orientation scores as channel representations -- 3.4 Multi-dimensional coding --

4. Channel-coded feature maps -- 4.1 Definition of channel-coded feature maps -- 4.2 The HOG descriptor as a CCFM -- 4.3 The SIFT descriptor as a CCFM -- 4.4 The SHOT descriptor as a CCFM --

5. CCFM decoding and visualization -- 5.1 Channel decoding -- 5.2 Decoding based on frame theory -- 5.3 Maximum entropy decoding -- 5.4 Relation to other de-featuring methods --

6. Probabilistic interpretation of channel representations -- 6.1 On the distribution of channel values -- 6.2 Comparing channel representations -- 6.3 Comparing using divergences -- 6.4 Uniformization and copula estimation --

7. Conclusions -- Bibliography -- Author's biography -- Index.

Abstract freely available; full-text restricted to subscribers or individual document purchasers.

Compendex

INSPEC

Google scholar

Google book search

This text collects a substantial amount of work on the topic of channel representations. Channel representations are a biologically motivated, wavelet-like approach to visual feature descriptors: they are local and compact, they form a computational framework, and the represented information can be reconstructed. The first property is shared with many histogram- and signature-based descriptors, the latter property with the related concept of population codes. In their unique combination of properties, channel representations become a visual Swiss army knife--they can be used for image enhancement, visual object tracking, as 2D and 3D descriptors, and for pose estimation. In the chapters of this text, the framework of channel representations will be introduced and its attributes will be elaborated, as well as further insight into its probabilistic modeling and algorithmic implementation will be given. Channel representations are a useful toolbox to represent visual information for machine learning, as they establish a generic way to compute popular descriptors such as HOG, SIFT, and SHOT. Even in an age of deep learning, they provide a good compromise between hand-designed descriptors and a-priori structureless feature spaces as seen in the layers of deep networks.

Also available in print.

Title from PDF title page (viewed on May 29, 2018).