Visual attributes

by Feris, Rogerio Schmidt Lampert, Christoph Parikh, Devi

Type Book

Rate 0.0 /5

Details

Files

Reviews

Download

Title Statements

Full title :Visual attributes / Rogerio Schmidt Feris, Christoph Lampert, Devi Parikh, editors.

Summary

This unique text/reference provides a detailed overview of the latest advances in machine learning and computer vision related to visual attributes, highlighting how this emerging field intersects with other disciplines, such as computational linguistics and human-machine interaction. Topics and features: Presents attribute-based methods for zero-shot classification, learning using privileged information, and methods for multi-task attribute learning Describes the concept of relative attributes, and examines the effectiveness of modeling relative attributes in image search applications Reviews state-of-the-art methods for estimation of human attributes, and describes their use in a range of different applications Discusses attempts to build a vocabulary of visual attributes Explores the connections between visual attributes and natural language Provides contributions from an international selection of world-renowned scientists, covering both theoretical aspects of visual attribute learning and practical computer vision applications This authoritative work is a must-read for all researchers interested in recognizing visual attributes and using them in real-world applications, and is accessible to the wider research community in visual and semantic understanding. Dr. Rogerio Schmidt Feris is a manager at IBM T.J. Watson Research Center, New York, USA, where he leads research in computer vision and machine learning. Dr. Christoph H. Lampert is a professor at the Institute of Science and Technology Austria, where he serves as the Principal Investigator of the Computer Vision and Machine Learning Group. Dr. Devi Parikh is an assistant professor in the School of Interactive Computing at Georgia Tech, USA, where she leads the Computer Vision Lab.

Edition Statements

Publication Date:2017.

Publication Statements

Publisher(s):Springer
Place of Publication:Cham
Language:eng
Publication Date:2017.

The Physical Object

Description:1 online resource : : illustrations

Edition Identifiers

ISBN:9783319500775, 3319500775
DOI:10.1007/978-3-319-50077-5
oclc:ocn978351031
LC Classification:BD352
Dewey Decimal:111

Preface; Contents; 1 Introduction to Visual Attributes; 1.1 Overview of the Chapters; References; Part I Attribute-Based Recognition; 2 An Embarrassingly Simple Approach to Zero-Shot Learning; 2.1 Introduction; 2.2 Related Work; 2.3 Embarrassingly Simple ZSL ; 2.3.1 Regularisation and Loss Function Choices ; 2.4 Risk Bounds; 2.4.1 Simple ZSL as a Domain Adaptation Problem ; 2.4.2 Risk Bounds for Domain Adaptation ; 2.5 Experiments; 2.5.1 Synthetic Experiments; 2.5.2 Real Data Experiments; 2.6 Discussion; References.

3 In the Era of Deep Convolutional Features: Are Attributes Still Useful Privileged Data?3.1 Introduction; 3.2 Related Work; 3.3 Learning Using Privileged Information; 3.3.1 Maximum-Margin Model 1: SVM+; 3.3.2 Maximum-Margin Model 2: Margin Transfer; 3.3.3 How Is Information Being Transferred?; 3.4 Experiments; 3.4.1 Object Recognition in Images; 3.4.2 Recognizing Easy from Hard Images; 3.5 Conclusion; References; 4 Divide, Share, and Conquer: Multi-task Attribute Learning with Selective Sharing; 4.1 Introduction; 4.2 Learning Decorrelated Attributes; 4.2.1 Approach.

4.2.2 Experiments and Results4.3 Learning Analogous Category-Sensitive Attributes; 4.3.1 Approach; 4.…

Subjects

Apprentissage automatique artificial intelligence Attribut (Philosophie) Attribute (Philosophy) Catégories (Philosophie) Categories (Philosophy) Computer vision Image processing Machine learning PHILOSOPHY -- Metaphysics User interface design & usability Vision par ordinateur