Dynamic semantic feature-based long-term cross-session learning approach to content-based image retrieval

Abstract

This paper proposes a novel content-based image retrieval technique, which facilitates short-term (intra-query) and long-term (inter-query) learning processes by integrating accumulated users' historical relevance feedback-based semantic knowledge. The history is efficiently represented as a dynamic semantic feature of the images. As such, the high-level semantic similarity measure can be dynamically adapted based on the semantic relevance derived from the dynamic semantic features. The short-term relevance feedback technique can benefit from long-term learning. Our extensive experiments show that the proposed system outperforms three peer systems in the context of both correct and erroneous relevance feedback.

Complementary relevance feedback-based content-based image retrieval

“…All the fixed parameters are the same as the ones used in [2]. & Xiao's dynamic semantic feature-based long-term cross-session learning system (i.e., DSF-based cross-session learning) [20]: This system uses the learned dynamic semantic features to measure the similarity among images. All the parameters can be automatically computed using the formulas specified in [20].…”

“…& Xiao's dynamic semantic feature-based long-term cross-session learning system (i.e., DSF-based cross-session learning) [20]: This system uses the learned dynamic semantic features to measure the similarity among images. All the parameters can be automatically computed using the formulas specified in [20]. Figure 5 shows a comparison among the average RP of the above eight CBIR systems using three kinds of NQs, which equal to 2 %, 5 %, and 10 % of the 6000 COREL benchmark images, to build the long-term repositories.…”

“…They are further used to estimate the semantic relevance between images. Xiao et al [20] propose a long-term cross-session learning scheme for CBIR. Based on the feedback log, they formulate the high-level dynamic semantic features for each database image in terms of retrieval patterns.…”

Block-based long-term content-based image retrieval using multiple features

“…Qi et al [6] apply the dynamic semantic clustering technique on the feedback log to cluster images into several semantic homogenous clusters for efficient retrieval. Xiao et al [7] use a dynamic semantic matrix to store and update users' feedback information in multiple sessions to capture more accurate semantic features. They then derive the semantic relevance among images and combine the semantic similarity with the visual similarity to find top images that are similar to the query image.…”

“…However, most existing RF techniques use shortterm (intra-query) learning to handle query formulation in a single retrieval session. Recently, long-term (inter-query) learning [2][3][4][5][6][7][8][9] extends short-term learning to derive semantic meaning of database images by studying the feedback history collected from multiple users in different retrieval sessions. Long-term learning has mainly been studied in two ways.…”