The following online demonstrations show the VISE software running on some public datasets. You can click on the links below to try visual search on these public datasets:

• Oxford Buildings: visual search on 5062 photographs containing several Oxford landmarks.
• 15cILLUSTRATION: visual search on 5130 images of illustrations printed in the 15th century (source: 15cI project).
• Scottish Chapbooks: visual search and grouping of Chapbooks printed in Scotland. (source: chapbooks project)

• User Guide
  • Installation and usage of VISE software is shown in this video recording of a Windows 10 desktop.
  • Installation and usage of VISE software is shown in this video recording of a MacOS-12.2 desktop.
• Frequently Asked Questions (FAQ)
• Users of VISE : lists research projects and commercial products that use the VISE software.
• Resources for developers
  • Source code repository of VISE
  • Compiling VISE from source in Linux and MacOS platforms.
  • Testing VISE using Oxford buildings dataset.
  • Specification of protocol to communicate with VISE server using HTTP REST API.
  • Details about the generic visual vocabulary used by VISE.

We manage support requests using our gitlab issues portal. If you want the support discussions to be confidential, tick the checkbox (below issue description) labelled This issue is confidential and should only be visible to team members. For the VISE software, we provide the following three types of support:

Individual User

Users can post their queries or report issues with the VISE software in our gitlab issues portal.

Research Support

We have supported many research projects (e.g. Humanities, History of Art, Literature, Book History, etc.) and individual researchers by provided all the required technical support in using the VISE software to answer novel research questions in their academic disciplines. We provide all the required technical support (e.g. installation, cleanup and indexing of large image datasets, web hosting of online search engine, etc.). Requests for research support can can be made by academic researchers from any country. Such requests should be posted in our issues portal with an issue label of Research Support.

Deployment Support

We provide technical support to institutions (e.g. museums, libraries, etc.) and commercial companies who want to deploy the VISE software in their own servers to power visual search of their personal image collections. The nature and extent of this support depends on the requirements of the end user. We also provide engineering support to upgrade the VISE software for specific requirements of an organisation. Requests for deployment support can be made by institutions or commercial companies from any country. Such requests should be posted in our issues portal with an issue label of Deployment Support.

The Users of VISE page contains a list of research projects and commercial products that use the VISE software.

VGG Image Search Engine (VISE) is a free and open source software for visual search of large collection of images using image region as a search query. VISE is released under a license that allows unrestricted use in academic research projects and commercial industrial applications. VISE is developed and maintained by the Visual Geometry Group (VGG) in the Department of Engineering Science of the Oxford University.

Development and maintenance of VISE software has been supported by the following two research grants:

• Visual AI: An Open World Interpretable Visual Transformer (UKRI Grant EP/T028572/1)
• Seebibyte: Visual Search for the Era of Big Data (EPSRC Grant EP/M013774/1)

VISE is developed and maintained by Dr. Abhishek Dutta, Dr. Relja Arandjelović and Prof. Andrew Zisserman.

Contact Abhishek Dutta for any queries or feedback related to this software application.

VISE source code can be download from its gitlab code repository. The VISE source is known to compile in GNU/Linux (gcc), Windows 10 (Visual C++) and MacOS (Clang).

The source code of VISE software is hosted at https://gitlab.com/vgg/vise. We want to nurture a vibrant open source community around the VISE software. Therefore, we encourage you to contribute and participate in the development of VISE. Our users can participate in the development of VISE software by reporting issues, contributing documentation, adding new features or improving existing features by sending a merge request.

VISE will be developed, maintained and supported by the Visual Geometry Group at least until November 2025.

If you use this software, please cite it as follows:.

Abhishek Dutta, Relja Arandjelović, and Andrew Zisserman. 2021. VGG Image Search Engine. Retrieved [INSERT DATE] from https://www.robots.ox.ac.uk/~vgg/software/vise/

  @misc{dutta2021vise,
    author    = {Dutta, Abhishek and Arandjelović, Relja and Zisserman, Andrew},
    title     = {VGG Image Search Engine},
    year      = 2021,
    url       = {https://www.robots.ox.ac.uk/~vgg/software/vise/},
    lastaccessed = {INSERT DATE}
  }