— Red Hen Rapid Annotator

Introduction

The Red Hen Rapid Annotator is a web-based application designed for rapid annotation of images and video.

Contact for queries: Peter Uhrig

For Rapid Annotator 2.0, click here.

Related scrolls

• How to use the NewsScape Edge search engine
• Manual tagging (with proposed Red Hen gesture tagging scheme)
• Integrating ELAN (development of export and import scripts)
• How to set up the iMotion annotator (draws rectangles on images to indicate event location)
• How to use the online tagging interface (integrated into Red Hen, but not frame accurate)
• How to use the Video Annotation Tool (online multi-dimensional video annotation interface for talks and demos)
• How to use the NewsScape Library search engine

Version 2 by Vaibhav Gupta

During GSoC 2018, Vaibhav Gupta created a new and improved version of the Rapid Annotator from scratch. Currently, a test installation can be found here (Beware, everyone has admin rights there). Please report bugs to Peter Uhrig. The source code can be found on Github. Some indications on how to use the system can be found in Vaibhav's GSoC blog.

[Anything below here is about the original version of the software.]

Rapid A nnotator1.0 by Peter Uhrig

Everything below this note refers to Version 1, whose use is now deprecated. But Version 1 still might be useful, and the description of its design and functionality remains useful for guiding the development of Version 2.

Image Annotation

The pisa server is used for image annotation.

Links and syntax

Rapid Annotator instructions (password protected):

• Instructions for students
• Instructions for experimenter

Annotation task syntax:

• http://YOURSERVER/annotator/annotator.html?dataset=sample&paramset=gender&user=testuser

For instance,

• http://164.67.183.180/annotator/annotator.html?imageset=MetMuseum&paramset=gender&user=kai (password protected)

To create different annotation jobs, set the imageset / dataset, the parameterset, and the user in the URL.

Note that the annotator will remember which images you have annotated and resume where you left off.

Image collections

• Politicians' emotion
  • clinton_smile1
  • cruz_smile1
  • mccain_smile1
  • obama_smile1
  • romney_smile1
  • sanders_smile1
  • trump_smile1
• Roman portraits
  • Ancient_Rome
  • Cambridge
  • MetMuseum
  • Monde_des_Cesars
  • Oxford
  • Roman_Coins
  • Rome_101
  • Zanker_I
  • Zanker_III
• test.collection

Annotation schemes

• emotion.param (smile, frown, neutral)
• figure.param (part of body represented in sculpture)
• gender.param (male, female, uncertain, etc.)
• material.param (precious stones and metals used in art objects)

Video annotation

The TVNews server is used for video annotation. The videos must be located with the NewsScape collection at UCLA. Access requires a UCLA Logon ID. Video annotation is under development and not fully documented.

Links and syntax

Rapid Annotator instructions:

• Instructions for students
• Instructions for experimenter

Annotation task syntax:

• http://YOURSERVER/rapid_annotator/annotator.html?dataset=sample&paramset=gender&user=testuser

For instance,

• https://tvnews.sscnet.ucla.edu/rapid_annotator/annotator.html?imageset=mindingthenews&paramset=gesture&user=test

To create different annotation jobs, set the imageset (in this case a list of videos), the parameterset, and the user in the URL.

The list of videos accepts these forms -- the UID, the file name (no extension), and the permalink:

477d7520-8e53-11e5-a327-0324b2e602dd 2904-2918 -- "Have I got news for you!"

2015-11-18_2200_US_WEWS_Live_on_5 2904-2918 -- "Have I got news for you!"

https://tvnews.sscnet.ucla.edu/edge/video,e2f3e252-005d-11e5-b05c-089e01ba0326,686 -- "Have I got news for you!"

The video clips will loop until a classification is selected.

Note that the annotator will remember which videos you have annotated and resume where you left off.

Development

The Red Hen Rapid Annotator was developed by Peter Uhrig in May 2016 for the rapid categorization of images. Support for video clips was added by Kai Chan at UCLA's Social Science Computing in November 2016. The code is available on Github, which is also where you report any issues.

• annotatorjs toolkit

.