Extraction of characteristic features of images with the help of the radon transform and its hardware implementation in terms of cellular automata

Abstract

Principles of implementation of cellular automata with hexagonal coverage are considered as applied to pattern recognition and image processing based on the Radon transform. Software and hardware implementations of the developed structure are simulated in the Active-HDL environment. The results of finding straight lines and segments and also of image recognition are presented.

Method of Hardware Selection of Characteristic Features Based on Radon Transformation and not Sensitive to Rotation, Shifting and Scale of the Input Images

“…5), and where for each image a graphical interpretation of the corresponding transformations is shown. [12][13][14][15][16][17][18][19][20][21][22][23] As can be seen from Fig. 5 -RT projection that is rotated on the 30° letter "A" in the direction 0° repeats completely the RT projection of not rotated letter "A" (0°) in the direction 30°.…”

“…But the above-mentioned angles (0°, 60° and 120°) exactly coincide with the directions of the hexagonal coverage ( fig. 1,а) [15,17]. Moreover, the hexagonal lattice is possible to allocate not only 3 but and 6 lines ( fig.…”

“…Stepan they form a parallel computing paradigm, just as Turing machine paradigm successive forms. Thus, they can be used as models which own natural parallelism [10][11][12][13][14][15][16][17][18].…”

“…In work [15], a comparison of the program models and VHDL-models is performed, and also the analysis of the received projections is carried out. The experiment confirmed the similarity of the projections built during software modeling and Active-HDL models.…”

Projection-Based Medical Image Compression for Telemedicine Applications

“…The periodic re-ordering of the elements of the Radon projections requires minimal interpolation and preserves all of the original image pixel intensities. Many researchers have proven the efficiency of Radon transform for face recognition [28], detection of finger prints [29], generation of unique nanotags for product anticounterfeiting [30], and extraction of characteristic features of images [31]. Performances of the proposed method are evaluated and compared with conventional and state-of-the art methods such as DCT, Haar wavelet, contourlet, and curvelet-based compression.…”