A Novel Image Retrieval Based on Visual Words Integration of SIFT and SURF

Area of Science:

• Computer Science
• Information Technology
• Digital Image Processing

Background:

• The exponential growth of digital image archives necessitates efficient retrieval methods.
• Content-Based Image Retrieval (CBIR) aims to bridge the semantic gap between low-level features and high-level image concepts.
• Existing CBIR methods face challenges in accurately representing complex visual information.

Purpose of the Study:

• To propose a novel approach for Content-Based Image Retrieval (CBIR) by integrating visual words from SIFT and SURF features.
• To enhance the robustness and accuracy of image retrieval systems.
• To address the semantic gap in image retrieval.

Main Methods:

• Integration of Scale Invariant Feature Transform (SIFT) and Speeded-Up Robust Features (SURF) visual words.
• Utilizing SIFT for robustness to scale and rotation changes.
• Employing SURF for robustness to illumination variations.

Main Results:

• The proposed integrated visual words approach demonstrated enhanced robustness in image retrieval.
• Qualitative and quantitative comparisons confirmed the effectiveness of the SIFT-SURF integration.
• Evaluations were conducted on diverse benchmark datasets including Corel and Oliva and Torralba.

Conclusions:

• The integration of SIFT and SURF features provides a more robust solution for Content-Based Image Retrieval (CBIR).
• This novel approach effectively addresses limitations of using single feature descriptors.
• The method shows significant promise for improving the performance of large-scale image retrieval systems.