AUTHORS: Stella Vetova, Ivo Draganov, Ivan Ivanov, Valeri Mladenov

Download as PDF

ABSTRACT: The paper below discusses the pros and cons of the local and global features in CBIR. To this end, four CBIR algorithms are designed and studied in terms of effectiveness. Two of them are based on local features extraction and the similarity is computed through Hausdorff distance or Euclidean distance respectively. The rest of the algorithms use global features extraction and the same two similarity distance metrics. For the feature extraction the Dual-Tree Complex Wavelet transform (DT CWT) is applied. The conducted experiments show that the local Features Algorithm with Hausdorff distance (LFAH) which was recently proposed in our previous study demonstrates better results in terms of effectiveness.

KEYWORDS: CBIR, DT CWT, image decomposition, global and local features, Hausdorff distance

REFERENCES:

[1] S. Banuchitra, Dr. K. Kungumaraj, “A Comprehensive Survey of Content Based Image Retrieval Technique,” Journal Of Engineering And Computer Science, vol. 5, issues 8, 2016 17584.

[2] V. Noutson, D. Argialas, P. Michalis, “Edge detection of manmade objects using wavelets in high resolution satellite images,” Annual Conference, 2007.

[3] G. Strang, T. Nguyen, Wavelets and Filter Banks, Wellesley-Cambridge Tress, 1998.

[4] H. Muller, W. Muller, D. S and T. Pun, “Performance Evaluation in Comtent-Based Image Retrieval: Overview and Proposals,” Pattern recognition Letters 22, issue 5, 2001, pp. 593-601. Precision and Recall ratio by image categories for the compared methods The implemented experiments highlight LFAH and most efficient from the tested ones using local and global features groups. It is concluded that LFAH requires a longer time period for image retrieval than GFAE. The reasons are founded on the executed opperations on image om the subimages and Hausdorff distance computation which needs a greater number of operations than Euclidean distance. The obtained average values as well as the curve show that LFAH demonstrates best ability for relevant image retrieval in the highest rank positions. Besides, according to a comparative analysis of four CBIR methods such as Local Histogram Method, Mean Segment Method, Variance Segment Method and the proposed LFAH, the last one generates most precise results. S. Banuchitra, Dr. K. Kungumaraj, “A Comprehensive Survey of Content Based Image Retrieval Technique,” International Journal Of Engineering And Computer , vol. 5, issues 8, 2016, pp. 17577- gialas, P. Michalis, “Edge detection of manmade objects using wavelets in high resolution satellite images,” ASPRS Wavelets and Filter Cambridge Tress, 1998. H. Muller, W. Muller, D. Squire, S. Maillet, and T. Pun, “Performance Evaluation in Based Image Retrieval: Overview and Pattern recognition Letters, vol. 601.

[5] G. Park, Y. Baek, H. Lee, “A Ranking Algorithm Using Dynamic Clustering for Content-Based Image Retrieval,” International Conference on Image and Video Retrieval, 2002, pp. 328-337.

[6] A. Berman, L. and Shappiro, “Efficient content-based retrieval: Experimental results,” CBA, 1999, pp. 55-61.

[7] U. Gargi, and R. Kasturi, “Image database querying using a multi-scale localized color representation,” CBA, 1999, pp. 28-32.

[8] I. Selesnick, R. Baraniuk, and N. Kingsbury, “The dual-tree complex wavelet transform”, IEEE Signal Processing Magazine, November, 2005, pp. 123-151.

[9] C. W. Shaffrey, N. G. Kingsbury, and I. H. Jermyn, “Unsupervised image segmentation via Markov trees and complex wavelets”, In Proceedings of IEEE International Conference on Image Processing, 2002, pp. 801-804.

[10] J. Magarey, and N. Kingsbury, “Motion estimation using a complex-valued wavelet transform”, IEEE Trans. on Signal Processing, special issue on wavelets and filter banks, vol. 46, No 4, April 1998, pp 1069-84.

[11] P. R. Hill, D. R. Bull, and C. N. Canagarajah, “Rotationally invariant texture features using the dual-tree complex wavelet transform”, in Proceedings of IEEE International Conference on Image Processing, 2002, pp. 901-904.

[12] B. Liao, and F. Peng, “Rotation-invariant texture features extraction using dual-tree complex wavelet transform”, International Conference on Information, Networking and Automation, 2010, pp. V1-361 - V1-364.

[13] S. Vetova, and I. Ivanov, “Image features extraction using the dual-tree complex wavelet transform”, 2 nd, International Conference on Mathematical, Computational and Statistical Sciences, 2014, pp. 277-282.

[14] W. Rasheed, G. Kang, J. Kang, J. Chun, and J. Park, “Sum of Values of Local Histograms for Image Retrieval,” Fourth International conference on Networking Computing and Advanced Information Management, 2008, pp. 690-694.

[15] J. Wang, J. Li, and G. Wiederhold, “SIMPLIcity: Semantics-sensitive Integrated matching for Picture Libaries,” IEEE Transaction on Pattern Analysis and Machine Intelligence, vol.23, no.9, 2001, pp. 947-963.

[16] J. Li, and J. Wang, “Autimatic Linquistic indexing of pictures by a statistical modeling approach,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 25, no. 9, 2003, pp. 1075-1088.

[17] W. Rasheed, Y. An, S. Pan, I. Jeong, J. Park, J. Kang, “Image Retrieval using Maximum Frequency of Local Histogram based Color Correlogram,” International Conference Multimedia and Ubiquitous Engineering, 2008, pp. 62-66.

[18] A. Daptardar,and J. Storer, “VQ based Image Retrieval using Color and Position Features,” Data Compression Conference, 2008, pp. 432- 441.

[19] M. Alsmadi, “An efficient similarity measure for content based image retrieval using memetic algorithm,” Egyptian Journal of Basic and Applied Sciences, vol. 4, issue 2, 2017, pp. 112-122.

[20] I. Sumana, M. Islam, D. Zhang, and G. Lu, “Content Based Image Retrieval Using Curvelet Transform,” IEEE 10th Workshop on Multimedia Signal Processing, 2008, pp. 11- 16.

[21] B. Bama, and S. Raju, “Fourier Based Rotation Invariant Texture Features for Content Based Image Retrieval,” National Conference on Communications (NCC), 2010.

[22] S. Chatzichristofis, Ch. Iakovidou, Y. Boutalis, and E. Angelopoulou, “Mean Normalized Retrieval Order (MNRO). A New ContentBased Image Retrieval Performance Measure,” International Journal “Multimedia Tools and Applications, vol. 70, issue 3, 2014, pp. 1767- 1798.

[23] H. Bhuravarjula, V. Kumar, “A Novel Content Based Image Retrieval using Variance Color Moment”, International Journal of Computer and Electronics Research, vol. 1, issue 3, 2012, pp. 93-99.

[24] S. Vetova, I. Ivanov, “Content-Based Image Retrieval Using The Dual-Tree Complex Wavelet Transform,” in Proceedings of the 2014 International Conference on Mathematics and Computers in Sciences and Industry (MCSI 2014), Varna, Bulgaria, September, 2014, pp. 165-170, DOI 10.1109/MCSI.2014.51.