This free survey is powered by QUESTIONPRO.COM
0%
 

Digital Image Retrieval Market Research Survey



You are invited to participate in our survey on the potential application of novel image retrieval technologies for use in our Management Studies degree market research project at Cambridge University Business School (www.jims.cam.ac.uk).



In this survey, approximately 20 stock photo archive company representatives will be asked to complete a questionnaire that asks questions about their current digital image handling process and the potential value that new technologies may offer their company. It will take approximately 5 minutes to complete the questionnaire.


Your participation in this study is completely voluntary. There are no foreseeable risks associated with this project. However, if you feel uncomfortable answering any questions, you can withdraw from the survey at any point. Furthermore, any answer can be left blank without affecting the progress of the survey. It is very important for us to learn your opinions.



Your survey responses will be stricly confidential and data from this research will be reported only in the aggregate. Your information will be coded and will remain confidential. If you have questions at any time about the survey or the procedures, you may contact Sean Moran by email at the email address specified below.


In return for your participation in this survey we would be happy to send you a copy of the final anonomised, aggregate results of the questionnaire. These results should be available after the 10/6/05. If you wish to know the results of the survey please do not hesitate to contact me at the e-mail address below.



Thank you very much for your time and support. Please start with the survey now by clicking on the Continue button below.


 
 

How many still images does your company currently hold?
 
0 - 1,000
 
1,000 - 10,000
 
10,000 - 100,000
 
100,000 - 1,000,000
 
1,000,000 - 10,000,000
 
10,000,000+
 
 

How much time does it typically take to find an image in your digital image archive that meets a customer's exact requirements (from customer request to retrieval)?
 
0-1 Minutes
 
1-5 Minutes
 
5-30 Minutes
 
30-60 Minutes
 
60+ Minutes
 
 
 
Briefly describe your process of image acquisition, image indexing and image archiving.
   
 
 
How important do you think the following aspects are when using an image retrieval system?



(Please note that in this question precision refers to the percentage of all relevant documents selected. Recall refers to the percentage of documents selected that are relevant.)



Not at all important Not very important Somewhat important Very important Extremely important
Accuracy of image content specification (e.g. 10% grass, 40% sky, water in bottom half of image only, people in center ... etc)
Search precision (at the expense of recall)
Search based on image content as opposed to an image label or caption
Cost of the new image retrieval system
Search recall (at the expense of precision)
Speed of search (possibly at the expense of reduced accuracy)
 
 
 
How can you tell when an image is out of date or unpopular (in terms of sales)?
   
 
 

Are your typical image search queries broad or precise?
 
Broad (e.g. scotch bourbon)
 
Precise (e.g (scotch OR bourbon) AND NOT (beer OR wine))
 
 
 
Please give an example of a typical query that you might use to search your image archive.
   
 
 
 
Please describe the nature of your current image dataset (e.g. wildlife images).
   
 
 

What are the major issues involved in meeting customer requests for images?
 
Difficult to find relevant images (e.g. irrelevant images are often retrieved)
 
Suitable queries hard to formulate
 
Image retrieval process too complex
 
Inflexible user interface
 
Other

 
 
 
Please describe the nature of your customer base (e.g. magazine companies).
   
 
 



Current academic research in the University of Cambridge Computer Laboratory is aimed at utilising state-of-the-art machine vision and artificial intelligence techniques to develop a more accurate and user-friendly method of searching and retrieving still images from large image archives.



Rather than simply relying on the image label or name (e.g. as does Google or Yahoo image search) to find relevant results, the technology analyzes actual image content in order to detect the presence of real-world objects (e.g. people, grass, clouds) and properties (e.g. texture, shape, colour) in images.



These objects and properties (combined with relations such as size and spatial proximity) can then be used in search queries to specify the characteristics of the desired images. For example, the prototype system would allow queries such as:



"some water in the bottom half which is surrounded by trees and grass, size at least 10%".



The central aim of the research is to provide image archivists with an intuitive and highly versatile means of expressing their retrieval requirements through the use of familiar natural language words and a straightforward syntax.



Crucially the speed of search is comparable to existing image retrieval technologies (less than 1 second, to a few seconds) and detailed manual labelling of images is NOT required. However, search based on traditional image labelling can work alongside the new system (i.e. it's not an either-or choice). Contextual knowledge available at processing time can also be used to satisfy a given user request.
 
 

Would this new technology add value to your current image retrieval process?
 
Yes
 
No
 
 
 
If No, please explain why not.
   
 
 
 
Would this technology save your company money? If so, how?
   
 
 
 
Would you foresee any problems with integrating the new technology into your existing image processing systems?

(Please ignore the technical details of the new technology and simply think of it as a generic bolt-on software package.)
   
 
Please contact [email protected] if you have any questions regarding this survey.
Survey Software Powered by QuestionPro Survey Software