What is the name of this problem in Computer Vision?

My google skills have failed me during research for a college project. Our task is write a program that describes an image in spacial terms, e.g. "Cat on the left of a chair", or for a simple image "Two triangles over a large circle, and a sqaure underneath it".

Suppose that after analysing an image I have a set of objects, with names, coordinates, sizes, orientations. I want to transform this set into sentences like above, i.e. describing relative positions, overlaying and so on.


Unfortunately, I cannot find anything on this subject. "Scene decription" means 3D graphics description languages. "Image description" is about the nouns and, recently, the verbs describing an image. It seems that I have the wrong keywords.

I'd much appreciate any hints about what to look for. Links to scientific papers to peruse (if you have any at hand) would be great too.

This seems to be exactly scene description as described, for example, in Timor Kadir's thesis:


Broadly speaking, the aim of scene description is to arrive at a set of descriptions of a real world scene which sufficiently capture the component parts of the scene, their positions, poses, motions and interactions. (p. 3)

StackExchange Computer Science Q#42394, answer score: 3