What is OCR?

Where the video/ screen recording transcripts transcribe what the respondents are saying, the OCR transcribes what the respondents are doing or showing you.

In effect, it transcribes the words/ text that appears on screen during a respondent's screen recording or video. 


This helps by:

  • Making the video searchable
  • Providing how long certain words/ strings of text were visible for during the recording
  • Allowing to use the AI View for analysis (for example, asking the AI to create a summary of a respondents online browsing journey by using the OCR data, which website did they visit and it what order etc) 


OCR is available for Moderators and Observers. It is not visible for Recruiters or Respondents.

How to use OCR


Step 1: Filter to the Project

This should be your first step when you first log in to Indeemo - this will ensure the other filters will work more effectively.

 



Step 2: Click on an Upload

When you click on an upload, the second tab will be 'Transcript' and the third tab will be ‘OCR’



Transcript: It is important to distinguish between Indeemo's Automated Video Transcripts and Indeemo's OCR.

Indeemo's transcripts show you what the respondents are saying. Indeemo's OCR shows you what the respondents are doing.

For more on the transcription, please see: Video Transcription: Editing a Transcript & Video Trimming


OCR: This tab is available on each Video and Screen recording. (This data can take a few minutes to generate, so may not be populated immediately when an upload arrives on the dashboard)



Step 3: Click on the OCR tab. 

OCR strings (data) are listed in order of the duration of appearance: so, words that appear for the longest amount of time on screen during the recording, will appear at the top. 


There are 3 columns of data:

1. The first column is the how long the string was visible for overall

2. The second column is the data: the string of text picked up on the screen.

3. The 3rd column is the point of appearance: the pink shapes indicate the point and duration of the text in the recording. 


You might notice some numbers, eg ’12.18’, this is the time from the clock on screen during a screen recording, picked up as text by the OCR.




You can scroll through this view, or you can hover over, or click on a pink shape:

  • Hovering over a pink shape shows you the timestamp it appears at
  • Clicking on a pink shape will take you directly to that point in the video where the string of text begins to appear. 


OCR Filter

It is important to distinguish between Indeemo's Keywords and OCR. 


Keywords: The keywords filter is scanning the respondents' transcripts, notes, captions or comments and listing the top 100 most frequently used words within your chosen search criteria (the other filters you have set).

You can view more on this feature here: Keywords


OCR: The OCR filter is scanning the OCR tabs on each video/ screen recording and listing the strings in order of what is seen on screen; the longest bring at the top within your chosen search criteria (the other filters you have set).

You can type into the search box in the filter to search for a string.



When you click on the Search icon to the right of the Filters, you will see on each upload and a number appear over the OCR button. 

This is the duration of the appearance of the filtered string(s) within this upload. 


Note: the occurrence of the strings can include overlaps in the video (example below).

Each string you select in the filter will be highlighted in orange in the OCR tab. 


OCR Word Cloud

The 'Word Cloud' gives 2 tabs – a Keywords cloud and OCR cloud.


Clicking on the OCR tab will show you a cloud of the 50 most visible OCR strings (data) (ensure to filter on your project first).

  • On the left, you will see a list of the strings of text along with the duration of time it appears. You can filter the OCR in this table based on duration or alphabetical order.
  • On the right you will see your OCR Cloud. 


 You can Export to Excel, or you can download the cloud image (this will generate a png.).




Using OCR in AI Analysis

For more on using the AI View for analysis, please see: AI Analysis