The Continuous Understanding Project

Home | People | Projects | Publications | Resources

Continuous Understanding Project | Sample Movie


On this page you will find instructions and manuals on how you can annotate the fruitcart corpus.

Specification Manual for version 19 can be read here. This manual is constantly being updated and can be regenerated from your anvil menu options: tools->create coding manual.

Click here, to download the latest specification file (NEW SPECIFICATION-29.) You can update the specification you are working with on anvil by opening your .anvil file with a text editor, and change the path

We listed some general rules for annotation. Here they are. If you feel we can add/modify some of them, please let me know.

You are welcome to go over more details on the specification scheme for annotation by reading this paper.


INSTRUCTIONS FOR SEGMENTING

Below you will find the .mp4 videos to be segmented according to the intervals given in the Segments column. Please follow this intervals since the video sound needs to match the transcripts in the .anvil files. From the .mp4, you must generate a .avi CINEPAK format with uncompressed speech. This is important. If you work in Windows, you can achieve this by using QuickTimePro. Mac users can segment videos with MPEG Streamclip.

After segmenting each interval, take a moment and verify that the .avi video you create matches the .anvil interval, the written transcript in the .anvil, and that the .avi video is playable in the Anvil program in your machine.

Naming your files. Please name the segmented files as follows: subjX-YYYY.avi, where X is the subject number (1-12), and YYYY is the start time interval.

Submitting your files. You can burn files onto a CD which I can give you or you can log in as a guest in the ftp I've set up in my local computer. Let me know which one you prefer and I'll give you more instructions on this.

Videos being segmented.
Subject Date given out Video Number Video Segments Anvil Transcripts Person in charge Comments Submitted
1 Nov 9, 2007 subj1 Kyle Audio and transcript Verified submitted subj2005101901
2 Nov 27, 2007 subj2 Chris Verified speech and video segments submmitted subj2005102401
3 Nov 27, 2007 subj3 Kyle Issue with overlapping intervals. Two segments missing Submitted subj2005102501
4 Dec 2, 2007 subj4 Kyle Submitted subj20051026002
5 Dec 2, 2007 subj5 Kyle Verified Submitted subj20051102001
6 Dec 2, 2007 subj6 Video Transcripts Chris Correct video has been found. Please make 4 segments according to the time intevals in transcripts and check consistency. subj20051109005
7 Dec 20, 2007 subj7 Chris Downloaded, checking transcripts submitted subj20051116001
8 Dec 20, 2007 subj8 Kyle submitted. subj20051121001
9 Feb 14, 2008 subj9 Video found. No transcripts found. subj20051026001
10 Nov 9, 2007 subj10 Dhairya Audio and transcripts Verified. Issues found for subj4 submitted subj20051122001
11 Abril 1 , 2008 subj11 Kyle Submitted
12 Abril 1 , 2008 subj12 Video Transcripts Kyle Last three segments do not match transcripts. In fact, they are the same video. Please check consistency with segments in the .anvil files. Video 48 is correct (see below). Video 49,50,51 are the same.
13 Abril 1 , 2008 subj13 transcrips with time aligned data found no video found


GENERAL INSTRUCTIONS FOR ANNOTATION

For training purposes, you can use the .anvil-example to guide your own annotation process. The .anvil-empty contains no annotation, only the time aligned transcript. If you find that the transcript and the video are not time aligned (i.e. the words do not match the video speech), please let me know before start annotating.

Moving a Track in Anvil. You can offset the transcript track, before annotation, by going to Edit->"Offset Track". There you enter the number of seconds (positive or negative) by which you want the track to be moved.

Naming your files. Please name your annotation files as follows: subjX-YYYY-URNAME.anvil, where X is the subject number (1-12), YYYY is the start time interval, and URNAME is your first name. You can email me the .anvil file.

Videos being annotated.
Video Date given out Subject/Segment .anvil file Video Comments Already submitted
Video 1 Oct 5, 2007 subj4 / 1625-2011 .anvil-empty, .anvil-example Video 1 Everyone annotates this video (training). Chase, Chris, Jake, Dhairya
Video 2 Oct 12, 2007 subj10 /1515-1917 .anvil-empty, .anvil-example Everyone annotates this video (training). Subject 10 Chase, Chris, Matt, Kyle, Jake, Dhairya,
Video 3 Oct 19, 2007 subj10 /4127-4523 Chris and Chase Chris, Chase
Video 4 Oct 26, 2007 subj10 /2402-2732 Chris and Stephen Chris, Stephen
Nov 2, 2007 No video given out. Catch up with annotation.
Video 6 Nov 9, 2007 subj10 /1034-1430 Chase and Matt Matt, 97%
Video 7 Nov 15, 2007 subj10 /1957-2314 Chris submitted
Video 8 Nov 15, 2007 subj10 /2823-3232 .anvil-empty Video Kyle (NEW!)
Video 9 Nov 15, 2007 subj10 /3311-3710 Matt Matt
Video 10 Nov 19, 2007 subj10 /3800-4047 Stephanie Stephanie
subj10 FINISHED
Video 11 Nov 27, 2007 subj1 /1414-2038 Matt Submitted/ but parts of annotation are missing
Video 12 March 18, 2008 subj1 /2148 Chris Chris
Video 13 Dec 6, 2007 subj1 /2736 Chris Chris- April 15, 2008
**** **** WINTER BREAK ****************
Video 14 Dec 20, 2007 subj1 / 3234-3647 Chris Chris
Video 15 Dec 20, 2007 subj1 / 3801-4324 Chris submitted
Video 16 April 5, 2008 subj1 / 4441-4926 Matt Matt
Video 17 Dec 20, 2007 subj1 / 5023-5359 .anvil-empty Video Chris (NEW!)
Video 18 Feb 1, 2008 subj1 / 5515-5920 Matt Matt
Video 19 Feb 1, 2008 subj2 / 2038-2429 .anvil-empty, video, Chris (NEW!)
Video 20 Feb 10, 2008 subj2 / 2522-2952 Chris
Video 21 April 10, 2008 subj3 / 1726 Matt, adjusted video .12 forward
Video 22 April 17, 2008 subj4 / 1625-2011 Chris training video
Video 23 April 22, 2008 subj5 / 3223-4040 Chris
Video 24 April 22, 2008 subj7 / 2129-2913 Matt
Video 25 May 26, 2008 subj7 / 3037-3620 Chris
Video 26 May 26, 2008 subj7 / 3713-4549 Chris
Video 27 May 26, 2008 subj7 / 4651-5215 Chris
Video 28 May 26, 2008 subj5 / 4221-4738 Chris Video not working...
Video 29 June 18, 2008 subj8 / 1918-2537 Chris Video has extra language at the beggining
Video 30 June 18, 2008 subj8 / 2624-2948 Chris
Video 31 June 18, 2008 subj8 / 3032-3318 Chris
Video 32 June 18, 2008 subj8 / 3401-3723 Chris
Video 33 July 2, 2008 subj2 / 3051-3434 Chris
Video 34 July 2, 2008 subj2 / 3532-3949 Chris
Video 35 July 2, 2008 subj2 / 4038-4504 Chris
Video 36 August 20, 2008 subj11 / 1557-2412 Chris
Video 37 August 20, 2008 subj 11 / 2458-2945 Chris
Video 38 August 20, 2008 subj 11 / 3030-3432 Chris
Video 39 August 20, 2008 subj 11 / 3535-4006 Chris
Video 40 August 20, 2008 subj 11 / 4104-4557 Chris
Video 41 August 20, 2008 subj 11 / 4640-5227 Chris
Video 42 August 20, 2008 subj 11 / 5322-5833 Matt
Video 43 August 20, 2008 subj 11 / 5930-6516 .anvil-empty, video, Matt Problems with video. Video is shorter than transcript.
Video 44 October 25, 2008 subj 12 / 1327-1700 Chris
Video 45 October 25, 2008 subj 12 / 1849-2217 Chris
Video 46 November 7, 2008 subj 12 / 2301-2634 .anvil-empty, video, Chris
Video 47 November 7, 2008 subj 12 / 2729-3059 .anvil-empty, video, Chris
Video 48 November 7, 2008 subj 12 / 3147-3447 .anvil-empty, video, Chris
Video 49 November 7, 2008 subj 12 / 3530-3845 .anvil-empty, video, Chris Video file does not match transcripts.
Video 50 November 7, 2008 subj 12 / 3947-421 .anvil-empty, video, Chris Video file does not match transcripts.
Video 51 November 7, 2008 subj 12 / 4307-4628 .anvil-empty, video, Matt -- Problem with videofile. Video file does not match transcripts.


Last change: 25 August 2008 by cgomez