A kind of feedback image retrieval algorithm based on PSO, Wavelet and subblock sorting thought
Kaiping Wei, Tingwen Lu, Bi Wu, Honghong Sheng · 2010
As the computer technology developed nowadays, image retrieval is developing fast. Due to the semantic disparity between low-level feature and high-level feature of an image. We propose a novel algorithm called a kind of feedback image retrieval algorithm based on PSO, Wavelet and sub-block sorting thought (RPWA) in the paper. Firstly, we preprocess the sample images and abstract the color histogram information as Fig.1. Secondly, extract n1 images randomly and their color information likes the first step. Thirdly, retrieve the similar images with PSO, if the retrieve results are satisfactory, output the images, else we use Wavelet to process the n1 images which are retrieved by the PSO, and re-initialize the particle swarm and continue to retrieve, we will get n3 images. Lastly, the low-frequency part and high-frequency part of the n3 images will be divided into 8 × 8 subblocks, the RPWA will use image sub-block sorting algorithm to determine the final result. In order to obtain more direct observation of results, I have developed two related systems. Experiments prove that the proposed algorithm is more accurate and effective.