| 摘要 |
PURPOSE:To completely retrieve images as the components of a scene expressed by an input sentence while using the sentence as the input of retrieval inquiry and to synthesize a lot of scenes even from a small scale image data base by integrating the image retrieving device into a system for editing and synthesizing natural images by arranging parts images on a background image. CONSTITUTION:This device is provided with an input means 101, language analyzing means 102, object identifying means 103, unfixed object decision rule 109, image selecting means 104, display means 105, word dictionary 108 and image data base 106 and further, a means is provided to retrieve an image by identifying whether objective data retrieving a language sentence inputted by the input means with the language analyzing means express the background or the parts corresponding to the unfixed object decision rule and the object identifying means and by dividing the objective data into the background and parts. |
