UAM ImageTool is a state-of-the-art environment for annotation of image corpora.

Version 2.0 now available! (10 october, 2010)

Powerful Annotation Tool

• Annotation of multiple images using the same annotation schemes, of your design.
• Annotation of each image at multiple levels (e.g., whole image, people, text elements).
• Searching for instances across levels, e.g., slogan in tobacco-advert, or beach-scene containing laura.
• Comparative statistics across subsets, e.g., contrasting mouth postures (smile, grim, etc.) across cultures, image types or genders.
• All annotation is in stored in XML files, meaning that your annotations can more easily be shared with other applications.
• We use "stand-off" XML: the annotation files do not contain the image, just pointers to the image. This allows for multiple overlapping analyses of the same image, not so easy in standard XML.

Easy to Use

UAM CorpusTool has been crafted to make the image annotation experience simple.

The Project Window is where you manage each project. It is used to add or remove layers from your study, to add or remove images to the corpus, and also to open each image for annotation at whatever layer.

Creating the annotation scheme is simple with the graphical scheme editor, allowing cross-classification as well as sub-specification of features. Any change in the scheme (e.g., deleting or renaming a feature) automatically updates all the annotation files in the project.

Annotating image segments is as easy as swiping a section of an image (click down and drag to end of section) and then indicating the features that are appropiate for that selection.
 

Searching your corpus is via the Corpus Search interface. Complex queries are possible, using features at different annotation levels (below we show eyes in female-face in cosmetic-advertisement).
 
You can also search for segments CONTAINING another segment type.
 

Statistical Studies: UAM CorpusTool supports a range of statistical analyses of your corpus, revealing patterns that the eye might miss. For instance, in the small study below we examine how people smile more in different advertisement genres.
 

Getting UAM ImageTool