FREE DM Review Site Registration!
Sign-up today and access DM Review on the Web!

Your FREE registration entitles you to:
FREE email newsletters

FREE access to all DM Review content

FREE access to web seminars, resource portals, our white paper library and more!

   

Content Management Mastered through Social Networking

Image-Seeker

  • DM Review Magazine, November 2007

REVIEWER: Ken Bouche, vice president of Trace for MyThings.com.

BACKGROUND:

MyThings.com is a global online registry of goods, where people can catalog and track their personal effects – everything from favorite gadgets to family heirlooms. The site also includes organizational features for keeping receipts and product information as well as social networking features for members to share information about their belongings if they wish. The company’s foundation is built upon identifying stolen property, especially art and antiques. The Web site began as a place where consumers and dealers would go to check if pending sales items were illegally obtained. Since then, MyThings.com has grown into two different areas. The first is MyThings, which maintains the registry of a person’s goods. The second part of the business, Trace, is the registry of known stolen goods. When an item is stolen, an individual, police official, insurance company or any other entity associated with the loss can upload an image of that object into the database. This sharing of information increases the opportunity to turn up leads or even solve the crime. Collectors, auction houses and other secondhand-market dealers search images on Trace.com to check if anything they may have purchased or are planning to sell has been reported stolen or lost. There is no charge associated with using the online search engine.

HARDWARE PLATFORMS: ASP.NET with Window 2003, IIS, MS SQL Server.

PROBLEM SOLVED: When images are provided regarding an item, Trace assigns both text search and visual matching functionalities to locate the image in future queries. LTU’s Image-Seeker is used to perform the visual matching – the search engine compares the uploaded picture with the existing database. Image-Seeker’s ability to search and capture images is invaluable to Trace’s users, particularly when serial numbers and text cannot be used for identification. Image-Seeker can search images without relying on the tags users attach to an image. Again, this makes it easier for people to find items that may not have all the necessary documentation or may be tagged with a file name that cannot be easily associated with the image. It is also important to note the significant impact LTU and Trace have made in the law enforcement arena focused on the identification of stolen goods. LTU’s ability to search the Trace database and match with images uploaded by users is helping police solve crimes in ways they never could before.

PRODUCT FUNCTIONALITY: Whenever an item is added to the application, its images are uploaded to Image-Seeker, which categorizes and indexes them. When a search is conducted a query containing the image in question is sent to Image-Seeker, which returns the best matches, ordered by the match probability.
There have been many success stories that illustrate the great benefits Image-Seeker has provided Trace’s users. For example, LTU’s Image-Seeker first identified a nineteenth-century statue that was about to be put up for sale at an auction house. The statue had been reported stolen in the mid-80s and had been entered into the Trace database by an insurance company’s claims department. This discovery not only allowed the item to be recovered and returned, it also prevented negative consequences for the auction house that could have later been associated with auctioning stolen items. Another example further demonstrates the value Image-Seeker and Trace brings to law enforcement. In cooperation with the Texas Police Chiefs Association and the Texas Department of Public Safety, MyThings is set to relaunch Trace with the statewide stolen articles files throughout the state. This integration will allow law enforcement officials to have direct access to Trace’s image bank of stolen items, providing greater detail and reach than is currently available in any of the 900 police departments’ own databases of stolen items.

STRENGTHS: Image-Seeker’s greatest strength is its ability to carry out dynamic search queries in a way that is superior to any other product available on the market. Because Image-Seeker can capture a wider number of results, the technology returns on matching correct images is greater thus increasing the number of stolen items recovered. With Image-Seeker, Trace is helping protect people and businesses alike from unknowingly buying stolen goods. This type of security and assurance is becoming more important as the economy grows and becomes increasingly dependent on the Internet market as a means for exchanging and selling second-hand goods. Trace is confident that with Image-Seeker, it has put itself in the best position for continued success and satisfaction among its users.

WEAKNESSES: The main weakness, with not only LTU Technologies, but with image recognition technology in general is that it still requires the human eye to intervene and make a decision on an exact match. Of course Trace would like to see the product take that last step and conclusively identify if two items are the exact same. However, the company is confident that it will happen; it’s just a matter of time.

SELECTION CRITERIA: The product was selected more than a year ago because it provided much better results than all other relevant products. More specifically, LTU’s solution has a broader engine and is able to identify photos and images in various settings rather than only finding exact matches. It allows for a more thorough search and provides a greater, more user-friendly opportunity to find the correct image.

DELIVERABLES: Trace uses the LTU engine in a very simple manner. An engine with an image file is the ‘input’ and then the ‘output’ is a list of images, from its database, which the engine identifies as best resembling the input image.
 
VENDOR SUPPORT: The way in which Trace is utilizing LTU is unique. Both organizations have worked together closely to change and customize Image-Seeker as needed – stretching it to its capacity. In addition, as new ideas or challenges have come up along the way, LTU has been tremendously responsive and enthusiastic about getting things completed.

DOCUMENTATION: LTU’s product documentation was quite straightforward and the team did not encounter any specific issues with implementation or since the solution has been up and running.

LTU Technologies: Image-Seeker
LTU Technologies
816 Connecticut Avenue, NW
Washington, DC 20006
(202) 558-4955
www.ltutech.com


For more information on related topics, visit the following channels:



Industry Vendors