Retrieval and Performance Evaluation of Artificially-Produced Images

N. Jamil, T.M.T. Sembok, and Z.A. Bakar (Malaysia)


Content-based image retrieval, performance evaluation, artificially-produced image, geometric shape descriptors, songket motifs


Artificially-produced images are stylized form of natural, real-life objects such as animals, flora or human. Motifs are repeated elements typically found in textile or architectural design. They are considered as artificially produced objects as they are customarily symbolic representations of natural objects. This paper describes and tests the efficiency of retrieving motifs using different combination of five geometric shape descriptors: eccentricity, compactness, convexity, rectangularity and solidity. Even though there are many existing approaches of shape-based image retrievals, the goal of this paper is to identify the fewest necessary shape descriptors to characterize the motif shape adequately so that it may be unambiguously retrieved or identified. Fifty selected songket motifs are used as sample queries in this paper. Similarities of these motifs are measured using Euclidean distance and performances of the retrievals are evaluated using both quantitative and qualitative tests. Quantitative test using recall-precision rate shows that geometric shape descriptors are effective in retrieving the songket motifs. However, qualitative test results demonstrate agreement between the system and the interviewed people in the assignment of similarity ranks is at average level.

Important Links:

Go Back