Automatic photograph retouching method. RU patent 2504840.

FIELD: information technology.

SUBSTANCE: automatic photograph retouching method involves creating a data array from photographs of different themes of classes, creating a database of references therefrom by interactive processing, based on Photoshop CS2, of predefined textures which give a comfortable perception of images of objects on the photographs, constructing a function of a photometric correction signal between the original and reference photographs, determining barcodes of the original photograph and the reference photograph by decoding brightness I(x,y) of matrices of images with the size |m×n| of elements in the matrix of intensities of tonal transitions with dimensions ||k×k|| of elements of the original and reference retouched photographs, algebraic subtraction of the matrix of barcodes while setting a threshold for positive identification of the reference, creating a reference address with barcode extension therefrom, a threshold difference and a photometric correction signal function, automatic search of the reference for the analysed photograph and retouching thereof based on the calculated barcode at the address in a reference database and the photometric correction signal function.

EFFECT: automatic search for the reference for a processed photograph by creating code features of each reference located in a reference database, and subsequent automatic photometric correction of the processed photograph on a reference texture mask stored in the reference database.

4 dwg

The invention relates to the computer engineering and may be used in systems of collection, conversion, processing of information in the various spheres of human activity: criminology, Astronautics.

Remote sensing of the Earth from space in the interests of the mineral exploration, forestry, ecology, monitoring of soil skin is performed by obtaining the digital images of the underlying surface. Selective characteristics of objects in images are color, tone, texture, topology. Environmental monitoring is carried out, as a rule, obtaining spectrozonal images on three sensitive layers of R, G, In ranges.

Known «Method of assessment of atmospheric pollution», Patent RU №2117286, 1998, A01G 23/00 - similar. Method-similar involves transformation of spectral brightness of the image I(x,y) in digital matrix |m n| elements in G, R parts of the visible spectrum, the by-element logical sorting of pixels in both matrices in accordance with algorithm: if R>G, R, if R<G, then R=R max -|k|·G, where k is the coefficient of correlation chromatic coefficients r, g, receive the result matrix of the same size, calculated numerical characteristics of the electrical signal output matrix: mathematical expectation, dispersion, envelope spatial spectrum, expect a histogram of pixels brightness, perform tie received relative distribution of the absolute values of the index of the atmosphere state in the region of its values and area control plots.

The disadvantage of analog is the dependence of the coefficient of correlation chromatic coefficients of r, g, shooting conditions, and the inadequacy of the algorithm of logical pixels sorting the physical process that distorts the resulting histogram of the distribution of pixels brightness and accuracy of the estimate of settlement parameter.

The closest analogue of the claimed technical solution is a method retouching images in interactive mode [see, for example, .., Processing graphics in Photoshop GS2, ed. «Eksmo», 2007, ch.7. «Image setup», page 170-186]. In the method, the nearest analogue of the region retouching set by its pre-selection, then carried filling a selection, the selected sight of the operator, certain shade of the many services offered by Photoshop predefined textures in the standard palette of colors BGR or in other palettes CMYK, Lab, giving a comfortable perception of the image of object.

Disadvantages closest you can consider:

- lack code characteristic that allows to automatically choose the reference picture for the image being processed;

- subjectivity comfortable perception of the image of the object, depending on the operator.

The task for the claimed invention consists in the automatic search benchmark for processed by creating a snapshot code signs for each sample, which is in the basis of the standards and subsequent automatic photometric correction processed snapshot mask texture reference stored in the standards.

Technical solution of the problem is achieved by way of automatic retouching of images includes dataset creation of images of different plot and classes, formation of the database of standards of them through processing «online» on the basis of specialized software, Photoshop CS2, pre-defined textures, giving a comfortable perception of the image of an object on the pictures, build a function photometric signal correction between the source and reference images, the definition of «barcodes» of the original picture and reference by recoding the brightness of I(x,y) matrices image size |mxn| elements in the matrix intensities of tonal transitions dimensionality ||kxk|| elements of the source and reference images, algebraic subtraction of matrices «barcodes» with the establishment of a threshold for reliable identification of the pattern formation of addresses of reference with the extension of his «bar code», the threshold difference and functions photometric signal correction, automatic search of reference for the analyzed image and retouching on the basis of calculated «barcode» at the basis of the standards and functions photometric signal correction.

The invention is illustrated by drawings, where:

figure 1 - functions signals photometric correction:

a) the brightness histogram of the pixels of the current 1 and 2 of the images;

b) standard (optimal) curve increasing the contrast of the majority of the snapshot closest analogue;

figure 2 - rendered images matrix barcode:

a) the reference snapshot, b) the original of the picture) algebraic subtracting matrices «barcodes»;

figure 3 functional diagram of the device, which implements a method;

figure 4 - the result of the automatic retouching of images:

a) analyzed the picture; b) automatically picture.

Technical essence of the method consists in the following.

Interactive improvement of the parameters of facsimiles widely applied in various a photo editor [see, for example, .., Processing graphics in Photoshop CS2, ed. «Eksmo», 2007, p.71-89, .145-151]. However, the level of automation of processes photo enhancement software is still very low: man-photo editor is forced to select optimal parameters of transformation of images in accordance with their artistic taste. Automation is to write special programs (scripts and actions) in the languages of the perceived photo editors, who can carry out a fixed sequence of transformations when you run the program.

While there are no universal methods of automatic optimization of the main parameters of the photos, because the selection of algorithms and parameters of such optimization depends on the plot of facsimiles. For example, menu General and the electoral contrasting images using low-frequency and high-frequency filtering closest analog is represented by a gallery, consisting of over 80 filters (see the nearest equivalent, .225-237).

Instruments interactive Refine your images are closest analogue:

- correction of histograms of distribution of pixels brightness, the command Histogram, Photoshop CS2, .171, illustrated in figure 1;

- improving contrast, the total lighten or darken the image, Levels, page 172;

- electoral change the contrast using the filter gallery, team Curves, the group Adjustment, the menu Filter, .174, 225-229.

In the proposed method for automatic retouching images use the code transformation matrix of the original image, brightness function I(x,y) dimension |m n| elements of the matrix barcode (matrix frames) size ||k x k|| the elements, where k is the maximum brightness value. Matrix frames is reversible transformation of the original image, but it carries the full information for the recognition of the plot of the image.

The algorithm of conversion includes the following procedures:

- ask matrix |k x k| elements «barcodes», where k is the maximum brightness value;

- choose window fixed aperture dimension of item 2 line;

- set the scan matrix within with 1 < i < m; 1 < j < n, where i is the number of rows, j is the number of the column;

- cyclically choosing two neighboring element x(i,j) and x(i,j+1) of the original matrix, and if the value of x(i,j)=a x(i,j+l)=b, then the matrix element «barcode» indexes (a, b) is increased by 1;

- display the elements of the matrix |k x k|a nonzero as nodes matrix barcode size of which is proportional to the accumulated values of k-th of brightness;

- calculate the diagonal element matrices «barcodes»

Listed procedures, as well as the algorithm of matrix subtraction «barcode» the current image from the pattern is implemented on the basis of specialized programs in Turbo Pascal.

The program of production matrix barcode.

The difference matrix (matrix difference can be obtained by simple algorithm

According to the results of the above operations form the address of the reference with the extension of his «bar code», the threshold difference and functions photometric signal correction for this class and plot analyzed images. Procedure of implementation of automatic retouching considered the example of a specific implementation.

Example of implementation of the method.

The claimed method can be implemented according to the scheme of figure 3. Functional diagram of the device figure 3 contains a flash-card 1, a set of client digital photos in one of the common formats (JPG, TIFF, RAW, and others), hardware-software system, Internet 2, implemented microprocessor based digital Converter 3, forming of digital facsimiles (client or reference) special matrix frames, images, RAM 4 containing a matrix frames client images, ROM 5, containing the master digital photographs in one of the common formats, ROM 6, which contains an array of matrices frames of reference photos implemented microprocessor based digital analyzer-discriminator 7, defines by pairwise comparison matrix frames for each client images nearest reference photo ID and a measure of dissimilarity between the client images from the reference implemented microprocessor based digital Converter 8, adjustment client digital images based on the measure differences of each client images from the corresponding reference, RAM 9, containing adjusted client digital photographs in one of the common formats. All components of the device are implemented on the existing technical infrastructure, PC such as Intel. Previously, in ROM 5 establish specialized software Photoshop QS2 with global access activated function commands photometric signal correction: Histogram, Levels, Curves. In RAM 4 establish a specialized program for calculation of the matrix barcode images.

Matrix barcode obtained on the image presented on the fig.2.

A visual image matrix barcode source image (Fig. 2A illustrates the fig.2.

Matrix barcode obtained from the original snapshot by using the «Curves». optimal photometric correction.

A visual image matrix barcode snapshot illustrates fig.2.

To establish the «conformity» of the original image to the standard of conduct algebraic subtraction of matrices with obtaining differential matrix. Measure compliance choose the establishment of a certain threshold criteria N<E, where N is the accepted norm of the matrix difference, such as the maximum of the absolute value, E is the number of installed expert. It is obvious that for images of different class and number of subjects (E) can vary within certain limits and must be installed experimentally. For the analyzed original snapshot fig.2 differential matrix looks like this.

The rendered image of differential matrix barcode illustrated fig.2. Database standards can be a relational database, field recordings which consist of pointer (address of the standard) in the structure of the matrix barcode, classifier (measures coincidence matrix differential) and expansion (team Curves - function photometric signal correction).

Efficiency of a method characterized by the possibility of automatic retouching of images with different plot with high reliability identification and quality of processing.

The way of automatic retouching the snapshot includes dataset creation of images of different plot and classes, formation of the database of standards of them through processing «online» on the basis of specialized software, Photoshop CS2, pre-defined textures, giving a comfortable perception of the image of an object on the pictures, build a function photometric signal correction between the source and reference images, the definition of «barcodes» of the original picture and reference by recoding the brightness of I(x,y) matrices image size |m n| elements in the matrix of the intensities of tonal transitions dimension ||k x k|| elements of the source and reference images, algebraic subtraction of matrices «barcodes» with the establishment of a threshold for reliable identification of the pattern formation of addresses of reference with the extension of his «bar code», the threshold difference and functions photometric signal correction, automatic search of reference for the analyzed image and retouching on the basis of calculated «barcode» at the basis of the standards and functions photometric signal correction.