SSH: Salient Structures Histogram for Content-Based Image Retrieval
Irfan Mehmood, Muhammad Sajjad, Khan Muhammad, Sung Wook Baik*
Key Figure

  • Access the paper
    • Proceedings of the 18th International Conference on Network-Based Information Systems, 2015 published [📃 Full-Text]
  • Additional Link
    •   IEEE Xplore
    • Abstract

      • Retrieving relevant information from large-scale image collections is a challenging task due to the complexity of visual patterns. Efficient indexing and retrieval mechanisms are essential for content-based image retrieval (CBIR) systems. This paper introduces a novel descriptor called Salient Structures Histogram (SSH), which captures salient image structures based on color and edge orientation features. Inspired by the human visual system’s sensitivity to these attributes, the proposed method represents these features as a sparse matrix and integrates them with saliency measures into a unified descriptor. The effectiveness of the SSH descriptor is evaluated on large image datasets, demonstrating improved discriminative capability and retrieval performance compared to existing state-of-the-art methods.

    • Additional Comments