Citations

Citing Video Clips

Through SOI’s TATOR, you are able to generate clips of existing dive videos and download them for your usage. When you generate and download a clip, an “Attribution” file will be associated with the video file. If you use content annotated through Schmidt Ocean Institute’s TATOR for publication, you need to properly cite usage of TATOR and usage of SOI footage. Please see below for instructions on how to cite TATOR.

Video citation format:

SOI Data Manager. ({year}). {video_id} [Video]. ROV SuBastian. Video, Imagery, and data provided by Schmidt Ocean Institute from expedition {cruise_id} accessed via TATOR. {URL}

Example:

SOI Data Manager. (2024). 2019-01-06 21:10:26+00:00.mp4 [Video]. ROV SuBastian. Video, Imagery, and data provided by Schmidt Ocean Institute from expedition FK190106 accessed via TATOR. https://cloud.tator.io/186/annotation/15643591?section=10446&sort_by=%24name&playQuality=720&playbackRate=1&frame=0&version=801&lock=0&fill_boxes=1&toggle_text=1

Citing TATOR Usage

In the case where you are not citing specific videos or frames but would like to cite the Tator platform, you can use the following format:

CVision AI [Computer software]. (2024). Retrieved from https://www.tator.io/

Citing Machine Learning Models

When using the outputs of ML models, you need to properly cite usage of such models. Given that all videos are processed through a set of baseline general-purpose object detectors, you should provide citations if you utilize the outputs of these models in your work.

In order to cite Megalodon, please follow the instructions provided here.

When citing the usage of an ML model, you should also discuss in your methodology how you utilize the outputs of the model. Please see below in Section 5 to learn about how TATOR harnesses these models to improve the accuracy and efficiency of the annotation experience. If you need any further information for your methodology section, please contact us at datasystems@schmidtocean.org