Use-cases

The volume of images and videos uncovered on a suspect computer during a forensic investigation can take hours to review.  The traditional step of running hash sets of known images against a computer is only effective for exact matches.  Using LTU finder, you are able to identify duplicate images (as you would with hash matching), but also modified matches of those images. 

Using finder during a forensic investigation allows you to first eliminate all known duplicate and modified versions of images.  These are known values, so there is no need for manual review.  Then, finder can provide suggested groupings for some of the remaining images - for example to highlight scanned documents or likely pornography.  These suggested groupings allow you to review thos images first, where you are most likely to find interesting evidence.

For videos, finder creates keyframe summaries of video files, greatly saving the investigator time over the traditional video process (carving the video files, finding the right codec, and watching the video.)

Content moderation is a complex matter. Social network, photo/video sharing sites, dating sites, but also auction sites, search engines, or mobile telecommunications networks, all have different rules to define if a piece of content is permissible or not. These rules are influenced by various factors: local culture, law, market positioning, corporate strategy,...

As a result, moderating or controlling the content created by a community is a heavy burden. It often requires significant human and computer resources to be performed well. Moderation resources within such an organization are critical to the success and the image of such sites. Our toolbox can help automate or speed up some tasks in the moderation process, reducing the grain of your moderation sift or improving the bandwidth of your moderation team.

Some studies estimate that up to 75% of images on the internet could be copyright infringements. It wouldn't be surprising in the sense that any image that ends up on a website and is not either a personal photo, a graphic file generated by the owner of the page or an image that has been acquired from a photobank, is a potential copyright infringement.

Looking for an image for your new product sheet, you incidentally found it on google images, or on another website. This is really the image you need, that will make your product sheet look best.

How to find where to buy it? Just go to a photobank equipped with LTU's visual searchbox, and upload your image to this box to find if this specific photobank has this image to sell, or a very simmilar one.

Still can't find your photo? Go to the LTU registry website and upload your image to look for it among the largest and most popular photobank's catalogs.

Competition among online photobanks is tough. It is largely based on how efficient search engines and the search experience are on each site. Here are examples where our image upload search brings photobanks a new search method that will generate additional revenue.

A web designer is looking for a very specific photo for the website he is currently working on - something very graphic that is hard to describe. He can't find anything good enough on any of the online search engines available at major photobanks. He finally ends up finding the image on Google Images, from a random website. This designer really needs that image, but he can't find who it belongs to or where to buy it.

Users of photo and video sharing and social networking websites often do not own the materials which they publish on the sites.

When the owners of  these images and videos object to their publication , the target of their objections is not directly the individual user who posted the content.  Whether liable or not, the immediate target is rather the platform of publication:   the websites that make the material available to the worldwide community.

Copyright owners understandably need copyright protection to protect revenues, including advertising revenues on the websites of the copyright owners.  And website operators need protection from unknowingly allowing the publication of copyrighted materials.

You know that the image is in there somewhere, but you just cannot find it.  The tags, the metadata you have searched with have not yielded what you were looking for.  And it is totally impossible to browse through all of the pages manually.

The dominant colors of an image - or the colors of the objects in an image - are sometimes one of the most important characteristics of the image.  LTU's visual search tools always have the ability to search for images based upon the colors of a query image, but the LTU engine also may be equipped refine image searches based upon a specified color.

After years working with major players in a wide array of different industry sectors, ranging from Patent and Trademarks Offices to Law Enforcement agencies or media companies, LTU has learned that large digital photo collections usually contain a significant amount of multiple copies of the same photo, that have been resized, reencoded or edited. This amount can range from less than 1% up to 20%.

Now, for a stock photo agency or a news agency, as well as for any organization storing photos or videos, the consequences of having twice the same photo or video in a collection can be a serious: