Content-Based Image Retrieval Using Features in Spatial and Frequency Domains

With the rapid increase of image data due to the development of information society, the necessity of image retrieval is increasing day by day. The research result of image retrieval using Local Binary Pattern (LBP) is reported previously, which is a robust feature obtained from spatial domain. It can be considered that we can get further information as the feature of image when extracting the feature from frequency domain. In this paper, we propose a novel image retrieval algorithm to improve retrieval accuracy, which using both features obtained from spatial and frequency domains. 2-dimensional Discrete Cosine Transform (DCT) is used to calculate the feature in frequency domain. Corel database is used for the evaluation of our proposed algorithm. It is demonstrated that image retrieval using combined features can achieve a much higher search success rate compared with that of algorithms using DCT and LBP, respectively. As a result, the precision rates and recall rates of this study was higher than the preceding study. In addition, better results were obtained using the weights.