Cogitum Digitizes Image Annotation Traditions

Challenge

Wouldn’t it be great if you could provide detailed information about every important element visualized on a digital image?

Technical illustrations, engineering drawings, product pictures, screen shots, aerial photos, design sketches, and artworks could become more valuable if you are able to record and share additional information about objects presented on them. These data could be used then for finding an image so search results would be more precise and relevant. As the number of images in your digital collection grows, ability to record comprehensive information about them that would help search and gain maximum knowledge becomes crucial. In corporate environment, this ability translates to time savings and more productive work.

Most tools and technologies currently available on the market lets you describe content of an image on the top level only. For example, you can assign a general description or list the names of the important objects as keywords. While the description can refer to the whole image only, no keyword can be clearly associated with the objects it describes.

However, providing ability to record information about individual elements is only half of the work.

In the today’s world of social networking and collaborative work, the workspace is not limited by someone’s desktop. Images are shared through e-mail, Web, and intranet so object-specific information about image content should always follow the image wherever it is moved.

Solution

Picture a scenario: you open an image in a special tool and place textual labels that describe individual elements or parts of the image directly on the picture. In each label you type a free explanation text, such as the name of the object, links to Web sites or other images, comments, questions, or suggestions. Since labels can have arbitrary location, clear identification of every important object is provided. You use this tool to collaboratively review images, document functionality, and suggest improvements.

As the labels are not imprinted into the picture, the original image remains intact. You should only click a button to hide the labels or make them visible. To find an image that contains a specific object, you just enter its name in a search field and the program shows you all the pictures where the element appears. When you need to share the image through email or Web with all the tags available, you just click another button and the image is sent or published in the Web with all object-specific labels preserved and optionally visible.

Sounds like a dream? Not for Cogitum LC, a Virginia-based company that develops content management solutions both for personal and professional use. Cogitum has developed FotoTagger, the image annotation technology and a series of tools that offer a standardized way to label image content on a low granular level, share it across communities in the Web, and find individual parts of image content.

While the most of image management tools traditionally consider the whole image a single and indivisible unit of content, FotoTagger enables describing any particular element shown on an image.

Like in the scenario described above, FotoTagger users place easy-to-hide annotation tags directly on an image and describe specific objects. As images are tagged, FotoTagger lets users easily find specific objects by their names or other text typed in the tags across piles of digital pictures. With FotoTagger, there is no need to use old-fashion techniques trying to meaningfully name the file or create a hierarchy of folders on a hard disk. A user can enter the name of an object and FotoTagger will find all pictures where this object is captured.

The illustration below demonstrates how an image annotated with the FotoTagger technology can look.

How It Works

To store and share annotation that describes individual elements or parts on an image, Cogitum developed FotoML, an open XML standard, that enables storing annotation, including text and position of each tag on the image.

A FotoML description is embedded into an ordinary JPEG file. On one hand, it is stored in the image file that allows keeping annotation tags available at any time wherever the image is placed – on a local computer, Web, or intranet – without having to compile distribution packages. On the other hand, the FotoML description is separated from image data so tags are retrieved and displayed over the picture rather than imprinting into the image itself.

This ability provides the following opportunities:

• Clear identification. Tags can be put directly on an image meaning clear identification of objects. Each tag can be positioned over the picture or even out of the image boundaries and connected to a corresponding element with a callout.



• Availability of original view. Being separated from the image data, tags are not merged with the picture. They can be hidden in a click of a button so the original view is always available. When the image is viewed in an application that does not recognizes the annotations format, they will be ignored and the original picture will be displayed.



• Searchability. It means that a tagged image can be found in the Web by text typed in the annotations. Unlike many photo sharing services that provide an ability to assign object-specific tags and allow finding images located on their own Websites only, pictures that contain FotoTagger tags can be found wherever they are located in the Web.



• Transferability. Tags always remain with the image file wherever it goes and are not lost when the image is shared or just moved to another location. No additional files should be distributed with the original image. FotoTagger enables direct publishing to Blogger.com, LiveJournal, and Flickr with annotation preserved.

Use of FotoML is invisible for end-users and does not impose any limitations, such as using specific vocabulary or following a specific procedure.

Learn more about how FotoML stores information about individual elements on the image.

To let image owners annotate individual objects on images and store annotations in FotoML, Cogitum offers software tools that enable labeling images located both in the Web and on local computers.

There are various products based on the FotoTagger technology:

• FotoTagger Desktop is a free software program that enables users to annotate images on a local computer. Annotated images can be published to Blogger.com or LiveJournal, as well as to a Flickr account. All annotations are preserved when sharing. Additionally, FotoTagger lets users to find an image by text typed in annotations that enables reaching the picture relying on meaningful information rather than on formal data, such as file creation date or location. FotoTagger can be downloaded for free from here



• FotoTagger Online Editor is a tool that allows you to annotate images collaboratively through the Web or intranet.



• FotoTagger Web API is a tool for Web developers that enable them to extract FotoTagger annotation notes from images and display them on Web pages. Website visitors will be able to see annotated images in the same way as they view them with the FotoTagger desktop tool. Annotation notes can be easily hidden to view an original image.



• FotoTagger Web Filter is a module that can be installed on a server in order to display images annotated with FotoTagger with easy-to-hide annotations. Website visitors will be able to see annotated images in the same way as they view them with the FotoTagger desktop tool.

Where It can be Used

Collaborative image annotation:

• Reviewing design sketches prepared by Marcom agencies online



• Discussing technical illustrations (screen shots, drawings, product photos, etc.) by technical writers and programmers/engineers



• Discussing home design by designers and clients

Image identification:

• Genealogy – family photos



• Customer support - explaining product capabilities on screenshots



• Fine art – commenting content on paintings

Send Us Your Feedback

We are always happy to get your feedback on the articles. It helps us constantly improve the content and provide you with the most relevant information.

If you know or want to know something that covered or should be mentioned in our articles, please feel free to contact us. We will be glad to get to know your opinion.