Content Based Image Retrieval (CBIR) is a popular method to search and retrieve the similar images. For medical applications, it plays an important role to find the diseasessuch as breast cancer in human body. Many existing methods were presented for improving the performance of CBIR method. Nevertheless, retrieval time and accuracy of CBIR are further to be improved. To solve this issue, an optimal classifier is to be used in CBIR. In this paper, Artificial Neural Network based on Particle Swarm Optimization based (PSO-ANN) is presented as an optimized classifier. Also, the features of images such as shape, texture, mean and standard deviation are extracted. To increase the speed of the classification, these extracted features are to be clustered using k-means clustering algorithm. From the clustered features, similar images of query image are retrieved using the proposed PSO-ANN classifier. Simulation results prove that performance of this proposed CBIR outperforms than that of existing methods in terms of accuracy and CBIR time.

基于内容的图像检索（CBIR）是一种搜索和检索相似图像的流行方法。对于医学应用，在人体中发现诸如乳腺癌的疾病起着重要的作用。提出了许多现有的方法来改善CBIR方法的性能。尽管如此，CBIR的检索时间和准确性仍需进一步提高。为了解决这个问题，将在CBIR中使用最佳分类器。本文提出了一种基于粒子群优化的人工神经网络（PSO-ANN）作为优化分类器。同样，提取图像的特征，例如形状，纹理，均值和标准偏差。为了提高分类速度，这些提取的特征将使用k-means聚类算法进行聚类。从群集功能中，使用建议的PSO-ANN分类器检索查询图像的相似图像。仿真结果证明，在准确性和CBIR时间方面，该CBIR的性能优于现有方法。