How to extract information about objects from an image, which is given in the form of a matrix of pixels? (The image is given as a pixel image, for example, because it was recorded by digital photography with a CCD.)
The extracted information can then for example be used for compression, because the information needed to encode an object, is at most the information from all points of the object with their corresponding colors (this coding is given). For example, for a green triangle displayed in the image, which consists of 200 green points, it is sufficient in the best case, to just represent the triangle and the points by a triangular object, for this only its three corners and the color is needed.
With the information about objects in the image, you can also to recognize objects at other (sub-) images. Or the information can be used to manipulate the image, e.g. to remove objects as a whole, copy or paste them.