V.V. Voronin, T.V. Morozova, M.M. Pismenskova

One of the main stages in many computer vision tasks is to search for similar blocks in the images and the video sequence frames. This task is relevant to the task of coding the motion vector in the construction of the blocks, tracing the trajectory of objects, reconstruction of areas lost two-dimensional signals, searching duplicate images, etc. Typically, as a block of selected small local area and brute force over the entire image or selection is determined by a similar unit in the Euclidean metric. This approach has a number of drawbacks. First, the search block bust leads to significant computational cost. Second, the Euclidean metric provides a cumulative assessment for the area. In this regard, it is interesting to develop other approaches to assessing the similarity of the image block. The different approaches to search for similar blocks in the images and the video sequence frames. The results of the comparison to search for similar units on test images with different geometric characteristics. Approaches to reduce the computational cost of algorithms based on the discrete cosine transform and hash functions. Provides a method for the synthesis of a similar blocks on the basis of a few blocks using texture analysis.

 

1. Voronin V.V., Morozova T.V., Makhov D.S., Samojjlin E.A., Minasjan D.G. Poisk lokalnykh osobennostejj na izobrazhenijakh s ispolzovaniem teksturnykh kharakteristik // Nelinejjnyjj mir. 2012. T. 10. № 11. S. 744-750.
2. Voronin V.V., Marchuk V.I., Pismenskova M.M., Morozova T.V. Sopostavlenie izobrazhenijj s otrazhajushhimi poverkhnostjami na osnove postroenija ASIFT-deskriptorov // Nelinejjnyjj mir. 2012. T. 10. № 11. S. 782-789.
3. Mood A., Graybill F., Boes D. Introduction to the Theory of Statistics (3rd ed.). McGraw-Hill. 1974. P. 229.
4. Gonsales R., Vuds R., EHddins S. Cifrovaja obrabotka izobrazhenijj v srede MATLAB. M.: Tekhnosfera. 2006.
5. Frantc V.A., Voronin V.V., Marchuck V.I., Egiazarian K.O. Fast texture and structure image reconstruction using the perceptual hash // Proc. SPIE 8655. Image Processing: Algorithms and Systems XI. 86550V. 2013.
6. Ojala T., Pietikäinen M., Harwood D. A comparative study of texture measures with classification based on feature distributions // Pattern Recognition. 1996. V. 29. P. 51-59.
7. Srinivasan G.N., Shobha G. Statistical texture analysis // Proceedings of world academy of science. 2008. V. 36. P. 1264-1269.
8. Laws K. Textured image segmentation. Ph. D. Dissertation. University of Southern California. 1980.