Content-based Image Browsing and Retrieval

We have developed innovative software to help users efficiently browse image collections and retrieve images of interest without the need for search terms. Our approach to browsing involves computer vision to extract descriptions from images, machine learning to adapt measures of image similarity based on user feedback, and a novel visualization method to arrange collections of images in ways that make browsing pleasing and efficient.

We have developed novel ways of describing images of objects such as textile design fragments; specifically, algorithms for analysis of regular visual textures (pattern repeats) present in such images and for obtaining image segmentation masks that approximate those used during colour printing.

Visualization automatically positions images within a large digital collection. Visual relationships between images are preserved while minimizing overlapping and occlusion of images. Placement of images is based on their visual content regardless of whether they share keywords.

We led the FABRIC project, a £1.4M collaboration with Liberty Art Fabrics, the V&A, and System Simulation Ltd., funded by the Technology Strategy Board. Browsing prototypes based on technology developed during this project have been evaluated with images from the Victoria & Albert Museum (V&A) for cultural heritage applications and from Liberty Art Fabrics for design inspiration within the commercial sector. The technology is also used in the Horizon interactive installation in the Compass lounge of the National Maritime Museum at Greenwich (a collaboration with the museum and three UK-based SMEs: Kin, Renderheads, and Liminal).

Publications
2014
[14] J. Han, S. J. McKenna,
"Lattice Estimation from Images of Patterns that Exhibit Translational Symmetry", Image and Vision Computing, vol. 32, 2014, pp. 64-73. [bib]
[13] J. Han, S. J. McKenna,
"Query-Dependent Metric Learning for Adaptive, Content-based Image Browsing and Retrieval", IET Image Processing (In Press), 2014. [bib]
2010
[12] R. Wang, S. J. McKenna, J. Han, A. A. Ward,
"Visualizing image collections using high-entropy layout distributions", IEEE Transactions on Multimedia, vol. 12, 2010, pp. 803-813. [bib] [pdf] [doi]
[11] R. Wang, S. J. McKenna,
"Gaussian process learning from order relationships using expectation propagation", in International Conference on Pattern Recognition (ICPR), 2010. [bib] [pdf] [doi]
[10] W. Jia, S. J. McKenna, A. A. Ward, K. J. Edwards,
"Classifying textile designs using region graphs", in British Machine Vision Conference (BMVC), 2010. [bib]
[9] W. Jia, S. J. McKenna,
"Classifying textile designs using bags of shapes", in International Conference on Pattern Recognition (ICPR), 2010. [bib]
2009
[8] A. A. Ward, A. M. Ward, J. Stephenson, S. J. McKenna, I. W. Ricketts,
"Art and Orphan Works: Ownership and Discovery", in Object and Identity in a Digital Age, Computers and the History of Art (CHArt) Conference, 2009. [bib]
[7] R. Wang, S. J. McKenna, J. Han,
"High-entropy layouts for content-based browsing and retrieval", in ACM International Conference on Image and Video Retrieval (CIVR), 2009. [bib]
[6] W. Jia, S. J. McKenna, A. A. Ward,
"Extracting printed designs and woven patterns from textile images", in International Conference on Computer Vision Theory and Applications (VISAPP), 2009, pp. 201-208. [bib] [pdf]
[5] J. Han, S. J. McKenna,
"Classifying and comparing regular textures for retrieval using texel geometry", in International Conference on Computer Vision Theory and Applications (VISAPP), 2009, pp. 347-354. [bib] [pdf]
[4] J. Han, S. J. McKenna, R. Wang,
"Learning query-dependent distance metrics for interactive image retrieval", in International Conference on Computer Vision Systems (ICVS), 2009. [bib] [pdf]
2008
[3] A. A. Ward, S. J. McKenna, I. W. Ricketts, R. Wang, J. Han, W. Jia, P. Sergeant, A. Buruma, P. Taylor, M. Stapleton, M. Selway, G. Howard, J. Stevenson, C. Wilkie, N. Sheen,
"Fashion and Apparel Browsing for Inspirational Content", in Multimedia Knowledge Management, MMKM Workshop on Future Directions, 2008. [bib] [pdf]
[2] A. A. Ward, S. J. McKenna, A. Buruma, P. Taylor, J. Han,
"Merging technology and users: applying image browsing to the fashion industry for design inspiration", in International Workshop on Content-based Multimedia Indexing (CBMI), 2008. [bib] [pdf]
[1] J. Han, S. J. McKenna, R. Wang,
"Regular texture analysis as statistical model selection", in European Conference on Computer Vision (ECCV), 2008. [bib] [pdf]
Powered by bibtexbrowser