Field Video Data Processing App

[thisdaria]

Summary

A python/flask/streamlit web application that assists in the data extraction/annotation step of field video processing!

If you are someone who has had to watch/document many hours of video data, have ideas about designing a user interface, or are generally interested in learning how to build an application, I would love to work with you! Ultimately, my goal for this week is to learn how to build a locally-hosted python web-app with a use case of interest!

Within the timeline of Ocean Hack Week, my hope is to build a minimum viable prototype with some of the following features:

• Video upload
• Page to input video metadata
• Page to define/upload configuration to dataset structure (column names/data types)
• Video display
• Play button
• Pause button
• Console next to video to add rows based on previosly defined data configuration to dataset based on video timestamp
• Page with live display/view of the data that has been recorded

Personnel

Daria Micovic
Anyone interested in this project, please reach out!

Specific tasks

• Get familiar with flask or streamlit and select
• Set up development environment
• Set up initial app framework
• Brainstorm/design/implement how columns/rows will be added to the dataset
• Write/generate documentation
• Practice software engineering best practices :-)

Data sets and infrastructure support

For the initial scope of this project, no external datasets or infrastructure support are needed and local machine resources will be sufficient for contributors! I have example video clips/data that we can use to get started which I will share with anyone interested!

The Motivation

Build a tool for use by those who process field video to automate some of the repetitive data recording.

Links to Explore

Below are some projects which are similar and have nice interfaces/consoles to get inspiration from.

• E-mammal
• Zooniverse
• VATIC
• Labelbox
• EventMeasure

[abkfenris]

I'd take a look at Panel or Streamlit as both will probably get you started faster than Flask.

[thisdaria]

I think within OHW that is probably what I'll start with!

[thisdaria]

(Streamlit)

[myrandaGoesToSpace]

Interested in this if it needs someone!