| 摘要 |
The invention concerns an image description method dissociating different description modes (1) of an image and separating the descriptions (3) into overlays (10) such that the superimposition thereof provides the complete description of the image. It is used to facilitate image referencing and search. The description of an image consists in determining a vocabulary contained in thesauruses (8) and a grammar (9) specific to the definition of said image, according to different image description modes (1), separately or jointly, and by dissociating the mental processes used during their description. With each of the description modes (1) is associated a description guide (9) which resorts to a vocabulary contained in thesauruses (8), to a grammar (9) and to a search and catalogue engine (7). The system enables to describe an image several times, according to several modes, to review or complete one or several description overlays (10) and to be performed by different persons. The images, the overlays, the vocabularies and grammar are stored in computer databases (8), (9), (10), (12), (13) and (15). The system can, among others, be used for referencing and searching for images in large image banks, via Internet. |
