i2m (Image to Model)

About i2m

The i2m (image-to-model) tool is web application that tells whether a given image is or is not a UML static diagram.

The tool is intended to classify as Positive those images that:

    Contain geometrical figures similar to those used in UML to represent classes, objects, notes, components, packages, ports, etc. (that is, rectangles or variations of rectangles).
    AND
    The figures contain textual information inside.
    AND
    The figures are linked to each other.
It is also intended to classify as Negative those images that do not satisfy the preceding rule, more specifically:
    The image satisfies the rule only partially, or only a certain portion of the image does.
    OR
    The image is not the product of a tool, but a manual drawing.
    OR
    The image is not a flat representation of a diagram (however, typical 3D effects or shadows to highlight flat figures are allowed).
    OR
    The image contains overlapping elements, ornamental elements, or any other kind of ambiguous elements.
Our experiments show a 95% success rate. You can help to improve the tool if you send us those images that you consider the tool classifies wrongly.
Contact: valentin.moreno AT kr.inf.uc3m.es

Go to Home to start.