| A key issue in content-based image retrieval (CBIR) research is exploring how to bridge the gap between the high-level semantics of an image and its lower-level properties, such as color, texture and shape. In this thesis, we present a new method using a set of perceptual features, called generic edge tokens (GET), as image shape content descriptors for CBIR. GETS represent basic types of distinguishable edge segments including both linear and nonlinear features, which are modeled as qualitative shape descriptors based on perceptual organization principles. In the method, an image is first transformed into GET map on the fly. The basic GETS can be grouped into higher-level perceptual shape structures (PSS) as well as additional shape descriptors. Image content is represented statistically by a set of perceptual feature histograms (PFHs) of GETs and PSSs. Similarity is evaluated by comparing the differences between the corresponding PFHs from the two images. Experimental results are provided to demonstrate the potential of the proposed method. |
