Leveraging Gaussian process approximations for rapid image overlay production

Research output: Chapter in Book/Report/Conference proceedingConference PaperResearchpeer-review

1 Citation (Scopus)

Abstract

Machine learning models trained using images can be used to generate image overlays by investigating which image areas contribute the most towards model outputs. A common approach used to accomplish this relies on blanking image regions using a sliding window and evaluating the change in model output. Unfortunately, this can be computationally expensive, as it requires numerous model evaluations. This paper shows that a Gaussian process approximation to this blanking approach produces outputs of similar quality, despite requiring signiicantly fewer model evaluations. This process is illustrated using a user-driven saliency generation problem. Here, pairwise image interest comparisons are used to infer underlying image interest and a Gaussian process model trained to predict the interest value of an image using image features extracted by a convolutional neural network. Interest overlays are generated by evaluating model change at blanking image regions selected using the prediction uncertainty of a Gaussian process regressor.

Original languageEnglish
Title of host publicationSAWACMMM'17 - Proceedings of the ACM Multimedia 2017 Workshop on South African Academic Participation
EditorsRiaan Wolhuter, Johan du Preez
Place of PublicationNew York NY USA
PublisherAssociation for Computing Machinery (ACM)
Pages21-26
Number of pages6
ISBN (Electronic)9781450355056
DOIs
Publication statusPublished - 2017
Externally publishedYes
EventWorkshop on South African Academic Participation 2017 - Mountain View, United States of America
Duration: 23 Oct 201723 Oct 2017
https://dl.acm.org/doi/proceedings/10.1145/3132711 (Proceedings)
http://www.sigmm.org/opentoc/SAWACMMM2017-TOC (Website)

Conference

ConferenceWorkshop on South African Academic Participation 2017
Abbreviated titleSAWACMMM 2017
CountryUnited States of America
CityMountain View
Period23/10/1723/10/17
Internet address

Keywords

  • Gaussian processes
  • Saliency generation

Cite this