Toward a Taxonomy and Computational Models of Abnormalities in Images

Authors

  • Babak Saleh Rutgers University
  • Ahmed Elgammal Rutgers University
  • Jacob Feldman Rutgers University
  • Ali Farhadi University of Washington

DOI:

https://doi.org/10.1609/aaai.v30i1.10468

Keywords:

Visual Attributes, Classification, Detection, Reasoning, Probabilistic Graphical Models, Crowdsourcing, Human Perception

Abstract

The human visual system can spot an abnormal image, and reason about what makes it strange. This task has not received enough attention in computer vision. In this paper we study various types of atypicalities in images in a more comprehensive way than has been done before. We propose a new dataset of abnormal images showing a wide range of atypicalities. We design human subject experiments to discover a coarse taxonomy of the reasons for abnormality. Our experiments reveal three major categories of abnormality: object-centric, scene-centric, and contextual. Based on this taxonomy, we propose a comprehensive computational model that can predict all different types of abnormality in images and outperform prior arts in abnormality recognition.

AAAI 2016 Proceedings Cover

Downloads

Published

2016-03-05

How to Cite

Saleh, B., Elgammal, A., Feldman, J., & Farhadi, A. (2016). Toward a Taxonomy and Computational Models of Abnormalities in Images. Proceedings of the AAAI Conference on Artificial Intelligence, 30(1). https://doi.org/10.1609/aaai.v30i1.10468

Issue

Vol. 30 No. 1 (2016): Thirtieth AAAI Conference on Artificial Intelligence

Section

Technical Papers: Vision