Similarity Retrieval Based on Spatial Inference for Image Database System

• Main Authors: Yi-Chu Lin, 林逸竹
• Other Authors: Dr. Yeong-Ren Jean
• Format: Others
• Language: zh-TW
• Published: 2011
• Online Access: http://ndltd.ncl.edu.tw/handle/23307235179941792399

碩士 === 靜宜大學 === 資訊管理學系 === 99 === There are many researches proposed in the spatial knowledge representation and spatial inference. Spatial knowledge representation is a technique of abstraction to describe spatial relationship between objects in an image. Spatial inference is a technique of similarity inference based on spatial relationships among image objects. In this paper, we provide a method to define the tuple of the distance and size between target object and reference object upon spatial inference. Making the spatial inference can represent image more precise. We also provide 48 similarity measure types to support flexible matching between the query picture and the database pictures for satisfying user’s different requirements. Finally, we use some examples to show that our image representation is more accurate than spatial inference method.