Ranking accurately collection images is the main objective of Content-based Image Retrieval (CBIR) systems. In fact, the set of images ranked at the first positions generally defines the effectiveness of provided search services, i.e., they are used for assessing automatically the quality of search systems as this set usually contains the collection images that are of interest. Recently, the use of ranking information (e.g., rank correlation) has been used in different research initiatives with the objective of improving the effectiveness of image retrieval tasks. This paper presents a broad rank correlation analysis for unsupervised distance learning on image retrieval tasks. Various well-known rank correlation measures are considered and two new measures are proposed. Several experiments were conducted considering various image datasets involving shape, color, and texture descriptors. Experimental results demonstrate that ranking information can be exploited for distance learning tasks successfully. Evaluated approaches yield better results in terms of effectiveness than various state-of-the-art algorithms.

The huge growth of image collections have demanded methods capable of conducting effective and efficient image searches. Among the most promising approaches, the Content-Based Image Retrieval (CBIR) systems have established as an alternative for automatically taking into account the visual information. Despite the important results achieved, retrieving relevant images (effectiveness) in minimal time (efficiency) remains a challenge task. Recently, unsupervised learning algorithms have been proposed to improve the effectiveness of CBIR systems by exploiting similarity and ranking information. Such algorithms does not require any user information, but often demand high computational efforts. On the other hand, parallel and heterogeneous approaches constitute a feasible solution for high performance computing. In this paper, we discuss a parallel and accelerated solution for computing the RL-Sim∗ Algorithm, a recently proposed unsupervised image re-ranking approach. The proposed algorithm uses the OpenCL standard, exploiting both CPU and GPU devives in an Accelerated Processing Unit (APU). The experimental evaluation demonstrated that significant speedups were achieved when compared with the original approach.