Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2018 Apr 25;13(4):e0194526.
doi: 10.1371/journal.pone.0194526. eCollection 2018.

An effective content-based image retrieval technique for image visuals representation based on the bag-of-visual-words model

Safia Jabeen  1 , Zahid Mehmood  1 , Toqeer Mahmood  2 , Tanzila Saba  3 , Amjad Rehman  4 , Muhammad Tariq Mahmood  5
Affiliations

An effective content-based image retrieval technique for image visuals representation based on the bag-of-visual-words model

Safia Jabeen et al. PLoS One. .

Abstract

For the last three decades, content-based image retrieval (CBIR) has been an active research area, representing a viable solution for retrieving similar images from an image repository. In this article, we propose a novel CBIR technique based on the visual words fusion of speeded-up robust features (SURF) and fast retina keypoint (FREAK) feature descriptors. SURF is a sparse descriptor whereas FREAK is a dense descriptor. Moreover, SURF is a scale and rotation-invariant descriptor that performs better in the case of repeatability, distinctiveness, and robustness. It is robust to noise, detection errors, geometric, and photometric deformations. It also performs better at low illumination within an image as compared to the FREAK descriptor. In contrast, FREAK is a retina-inspired speedy descriptor that performs better for classification-based problems as compared to the SURF descriptor. Experimental results show that the proposed technique based on the visual words fusion of SURF-FREAK descriptors combines the features of both descriptors and resolves the aforementioned issues. The qualitative and quantitative analysis performed on three image collections, namely Corel-1000, Corel-1500, and Caltech-256, shows that proposed technique based on visual words fusion significantly improved the performance of the CBIR as compared to the feature fusion of both descriptors and state-of-the-art image retrieval techniques.

PubMed Disclaimer

Conflict of interest statement

Competing Interests: The authors have declared that no competing interests exist.

Figures

Fig 1
Fig 1
(a) Semantic gap-Corel images of two different semantic categories (i.e. “Mountains” and “Beach”) with close visual appearance; (b) Two sample images of different shapes with close visual and semantic appearance (images used in the figure are similar but not identical to the original images used in the study due to copyright issue, and is therefore for illustrative purposes only).
Fig 2
Fig 2. Methodology of the BoVW based image representation for CBIR.
Fig 3
Fig 3. Block diagram of the proposed technique based on visual words fusion (images used in the figure are similar but not identical to the original images used in the study due to copyright issue, and is therefore for illustrative purposes only).
Fig 4
Fig 4. Sample of images from different semantic categories of the Corel-1000 and Corel-1500 image collections (images used in the figure are similar but not identical to the original images used in the study due to copyright issue, and is therefore for illustrative purposes only).
Fig 5
Fig 5. Performance comparisons of standalone SURF, standalone FREAK, and features fusion of SURF and FREAK descriptors techniques on different sizes of the dictionary for the Corel-1000 image collection.
Fig 6
Fig 6. Comparison between visual words fusion vs. features fusion of SURF-FREAK using proposed technique on the Corel-1000 image collection.
Fig 7
Fig 7. Performance comparison in terms of PR-curve on the Corel-1000 image collection.
Fig 8
Fig 8. Image retrieval result shows a reduction of the semantic gap using automatic image annotation on the semantic category “Flowers” of the Corel-1000 image collection (images used in the figure are similar but not identical to the original images used in the study due to copyright issue, and is therefore for illustrative purposes only).
Fig 9
Fig 9. Image retrieval result shows a reduction of the semantic gap in the semantic category “Horses” of the Corel-1000 image collection (images used in the figure are similar but not identical to the original images used in the study due to copyright issue, and is therefore for illustrative purposes only).
Fig 10
Fig 10. Performance comparisons of standalone SURF, standalone FREAK, and features fusion of SURF-FREAK techniques on different sizes of the dictionary for the Corel-1500 image collection.
Fig 11
Fig 11. MAP performance comparison of the proposed technique based on visual words fusion vs. features fusion of SURF-FREAK technique on different sizes of the dictionary for the Corel-1500 image collection.
Fig 12
Fig 12. Performance comparison in terms of PR-curve on the Corel-1500 image collection.
Fig 13
Fig 13. The retrieved images show a reduction of the semantic gap in response to the query image taken from the semantic category “Tigers” of the Corel-1500 image collection (images used in the figure are similar but not identical to the original images used in the study due to copyright issue, and is therefore for illustrative purposes only).
Fig 14
Fig 14. Performance comparisons of standalone SURF, standalone FREAK, and features fusion of SURF-FREAK techniques on different sizes of the dictionary for the Caltech-256 image collection.
Fig 15
Fig 15. MAP performance comparison of the proposed technique based on visual words fusion vs. features fusion of SURF-FREAK techniques on different sizes of the dictionary for the Caltech-256 image collection.
Fig 16
Fig 16. Performance comparison in terms of PR-curve on the Caltech-256 image collection.
See this image and copyright information in PMC

References

    1. Qiu G., Color image indexing using BTC. IEEE Transactions on Image Processing, 2003. 12(1): p. 93–101. doi: 10.1109/TIP.2002.807356 - DOI - PubMed
    1. Yu F.-X., Luo H., and Lu Z.-M., Colour image retrieval using pattern co-occurrence matrices based on BTC and VQ. Electronics letters, 2011. 47(2): p. 100–101.
    1. Lu Z.-M. and Burkhardt H., Colour image retrieval based on DCT-domain vector quantisation index histograms. Electronics Letters, 2005. 41(17): p. 956–957.
    1. ElAlami M.E., A novel image retrieval model based on the most relevant features. Knowledge-Based Systems, 2011. 24(1): p. 23–32.
    1. Guo J.-M., Prasetyo H., and Su H.-S., Image indexing using the color and bit pattern feature fusion. Journal of Visual Communication and Image Representation, 2013. 24(8): p. 1360–1379.

Publication types

MeSH terms

LinkOut - more resources