2020 ESA Annual Meeting (August 3 - 6)

COS 215 Abstract - View eXtract aNnotate (ViXeN) media files: A project manager for multimedia data

Kadambari Devarajan, Organismic and Evolutionary Biology, University of Massachusetts at Amherst, Amherst, MA; Environmental Conservation, University of Massachusetts at Amherst, Amherst, MA and Prabhu Ramachandran, Aerospace Engineering, Indian Institute of Technology Bombay, Mumbai, India
Background/Question/Methods

Multimedia data, such as images and videos from trail cameras, and audio files from recorders of amphibian and bird calls, have become integral to monitoring and conserving wildlife. However, there is a dearth of general-purpose computational tools to access and process such multimedia data, especially those conducive to research involving more than one type of media. ViXeN (https://vixen.readthedocs.io), an acronym for View eXtract aNnotate media data, is a simple, free and open-source media metadata management tool that facilitates managing the metadata associated with media files such as camera trap videos and images. It is general purpose and extensible, with support for images, videos, audio files, text, and portable document format (PDF) files.

This media manager was initially designed to handle several thousand camera trap videos for an ecological study involving mammalian carnivores and is capable of handling large datasets (of the order of a million files). ViXeN has subsequently been used to conduct a systematic literature review as well as general-purpose photograph management, and can be applied to a variety of projects involving media files. It is written in Python and run as a web browser-based user interface (UI) to view the media and simultaneously edit the metadata tags, making it intuitive, convenient, and portable, while supporting several file formats. It is non-intrusive and allows researchers to associate any number of arbitrary metadata tags with the media. The package allows us to export these tags into a comma separated value (CSV) file for further processing and analysis that can be done with existing software and programming languages the user is comfortable using.

Results/Conclusions

The ViXeN toolkit supports easy import and export of data, thereby facilitating data analysis. By virtue of its design, data entry errors are minimized while significantly accelerating the entry process. ViXeN allows a user to run processing scripts on the media files and can also be used to automate the extraction of tags. Furthermore, the complex search and processor functionality provided by ViXeN offer scope for adding features such as automating identification and batch processing media files. It is easy to install and currently available for Linux, Windows, and macOS (https://github.com/vixen-project/vixen).