Running Algorithms in Portal

The algorithms in portal are listed in more detail in the Algorithm Library. This page aims to give an overview of where to launch algorithms, and how to find the results.

Where to Launch Algorithms

My Dashboard > Video & Images

The landing page for the Portal can be returned to at any time by clicking the logo or the header link to "My Dashboard". Selecting a project highlights the Video & Images view of your selected project or folder.

The media card can be clicked in the center to visit the player page, or on top right to select the file. Additionally you can select the entire view (based on the pagination, ie. 15, 30, 50, 100) and the available actions will be presented to the user. One is Run Analysis with AI. Choosing this option will open a popup with any available algorithms. Select the algorithm to meet your needs and hit apply. Currently algorithm results overwrite the previous runs. In the future you can preserve more than one algorithm run.

Project or Folder > Analysis tab

Each folder within a project has an Analysis tab which will show a summary for completed algorithms. It also allows you to launch the algorithm using the same modal window but apply it to the entire folder.

Media > Analysis

Media > Analysis View

At the Media-level there are two places where you can launch analysis with AI. The first the media-level Analysis page. You can get there from your dashboard by clicking the three dots next to the file name and clicking Analysis, by clicking the analysis text on the card, or by clicking into the tab from another media specific view like the player view.

Media > Player View

From the player view there is an option to launch an algorithm from this page, or to go to the Media > Analysis tab to launch it as described in the previous section.

Glossary

Algorithm - Refers to the combined toolage we use to generate information with less human input. Portal algorithms use fine-tuned metrics from specific Models to determine regions of interests, and add a Label if it is suggested.

Weights - Training data creates fine tuning for specific use cases for each named algorithm in our system. We currently offer two MBARI-trained weights added via a model submission form on the FathomNet Portal website by ML Ops team.

Model - Can be used interchangeably with algorithm. Multiple models which have specific use cases and strengths.

Detection - A region of interest in a video frame or image. Algorithms may be “detection only” meaning all results will have “No Label” or be generically labelled as “object.”

Classification - Assignment of a label to an Observation by an algorithm. Algorithms may be “classification only” meaning they suggest the content, but won’t box it. A trained classification and detection algorithm is ideal in that it does both.