Cortex
COYOTE
https://github.com/coyote-team/coyote
https://vimeo.com/206662220/b9fe9be393
Coyote is a project developed by cultural heritage professionals and people from the accessibility community to encourage the use of visual description in museum practice. Originally intended to help people who are blind or have low vision experience images online, the project’s scope has expanded to encompass other uses for visual description, to serve both sighted visitors and those with impaired vision. The Coyote software—an open-source, cloud-hosted toolkit developed by the project team—helps organizations with the workflow around creating, reviewing, managing, and publishing descriptions.
Coyote is a web application implemented with standard web 2.0 technologies. The frontend is comprised of an html5 interface with JavaScript and CSS being used to achieve an accessible and dynamic user interface. With respect to accessibility, WCAG 2.0 AA guidelines have been targeted, however, given that the very nature of Coyote involves unlabeled images, not all functional criteria are appropriate. The backend is implemented on top of the popular Ruby on Rails platform. Coyote both consumes and offers RESTful APIs. External environments such as websites, mobile apps, or other software can request image descriptions from Coyote by using said RESTful API. Coyote learns about the images needing to be described by consuming external APIs such as that of a CMS. While Coyote’s primary goal is to be the canonical holder of image description, it does not strive nor is meant to be the canonical holder of image data or metadata not related to description. Coyote offers role-based logins to facilitate a workflow approach towards image description. To this end, those who are administrators may approve, assign, and review descriptions. When users log in, they are presented with a queue-based interface for authoring descriptions that they have been assigned. Realizing that image description is a context-sensitive process, Coyote offers content creators the ability to author both multiple types and multiple instances of a single type of description per image. In addition, a tagging interface exists to assist in search and organization activities within the interface.