Adobe Photoshop is the world's most popular photo editor for creating really cool things. Today you will learn how to get only its outline from an image. This can be useful, for example, for creating a coloring book for a child.

Simple drawings that do not have intricate detail are easiest to trace using the Pen tool. It will be faster and easier this way. With images that have more complex details, the scope of work is somewhat different.

Step-by-step instruction

1. Load the original image into Photoshop.

2. Now let's start working with filters. Go to the menu "Filter" - "Stylize" - "Select edges".


The image will look like this:


3. Now also open the "Filter" menu and go to "Sketch" - "Photocopy". A window will open, in the right part of which you need to set the following parameters: "Detailing" - 9; Darkening - 5. Click OK. Note: it is not necessary to strictly follow the instructions at this point. Experiment with the detail and shade values ​​to get the look that works best for you.


4. Go to Image> Adjustments> Brightness / Contrast and adjust the sliders until you get the best results.




Save the finished image to your computer in JPEG format. If you created a coloring book for a child, just print the image on a printer.

Monochrome, outline image

First letter "c"

Second letter "and"

Third letter "l"

The last beech letter "t"

The answer to the question "One-color, contour image", 6 letters:
silhouette

Alternative crossword questions for the word silhouette

Face contour

m. fr. shot from the shadow, from the side outline of the face

Poem by M. Lermontov

Image, outline

Cut out outline of the subject

Definition of silhouette in dictionaries

Explanatory dictionary of the Russian language. D.N. Ushakov Meaning of the word in the dictionary Explanatory dictionary of the Russian language. D.N. Ushakov
silhouette, m. One-color contour image of a human subject against a background of a different color, drawn or cut out. transfer Vague outlines of something, visible in the darkness, fog. Here lights flashed, silhouettes of huts. Chekhov. From time to time ...

Wikipedia Definition of a word in the Wikipedia dictionary
Silhouette is one of the islands of the Seychelles archipelago. Located in the Indian Ocean, belongs to the state of Seychelles.

Explanatory Dictionary of the Living Great Russian Language, Dal Vladimir Definition of the word in the dictionary Explanatory Dictionary of the Living Great Russian Language, Vladimir Dal
m. fr. shot from the shadow, from the lateral outline of the face.

Explanatory dictionary of the Russian language. S.I.Ozhegov, N.Yu.Shvedova. Meaning of the word in the dictionary Explanatory dictionary of the Russian language. S.I.Ozhegov, N.Yu.Shvedova.
-a, m. One-color planar image of an object against a background of a different color. S. face in profile. transfer Outlines of something, visible in the darkness, fog. S. of the mountain range. Lines, contour of clothes. Fashionable s. clothes. adj. silhouette, th, th.

Examples of the use of the word silhouette in literature.

Fighters began to interact more clearly with anti-aircraft artillery, they operated at heights inaccessible to artillery, used a light background above the target, created by luminous bombs, and tracked against this background silhouettes our aircraft, gave a signal to the anti-aircraft gunners to cease fire and went on the attack.

In the direction of Anapa, against the background of clouds, were already visible silhouettes heavy aircraft.

An arrow whistled just above his ear, the crossbowman unloaded his weapon into the spiral staircase that appeared on the silhouette- the magician has already raised his hands, preparing to cast a spell.

Senior Lieutenant Arsenyev looked up from the periscope and rubbed his eyes: he dreamed of some kind of lights and dark silhouettes ships, but he was immediately convinced of the error.

The creatures disembarking from the ships were beyond imagination with their silhouettes shaped like coils of a spiral or blossoming arum flowers, with bodies of purple color and with heads resembling starfish.

Drawing on points for children of lines, figures and animals. Draw point by point to develop writing skills.

A beautiful handwriting and successful learning to write depends on the correct use of a pencil, skillful pressure and the ability to draw lines of all kinds of shapes. Start by learning to draw lines and shapes by points, and then have your child draw and color animals by points.

Draw point by point, developing skills gradually

Drawing lines with a pencil or pen is a great practice to help train your hand to write, develop small muscles, and teach your baby to hold something firmly.

The dotted line serves as a guide and helps the child, because at any time you can slow down the speed of drawing, increase or decrease the pressure on the pencil, without spoiling the picture, and, therefore, without losing interest.

As soon as the child learns to draw lines by points, straight and all kinds of waves, move on to the figures, and then to the animals. Bending the dotted lines will develop your drawing skill enough to start learning the spelling of letters and numbers.

When offering your child a printed material with a picture on which you want to draw something point by point, first ask the child to trace the lines with the index finger of his right hand (or left, if the child is left-handed). Then ask him to draw with his finger, not on the sheet, but as if in the air above the picture. Repeat the exercise several times, and then complete the task with a pencil.

When your child learns to draw point by point with a pencil, offer him a pen or marker.

Pay attention to drawing on the points of the animals, without taking your hand off the paper.

How else to develop fine motor skills, besides drawing by points?

If your child for some reason is not interested in the point-by-point drawing materials, you can have fun developing fine motor skills in other ways.

1. String together large beads on strings or sort the beads;
2. Glue a large sheet of paper or old wallpaper on the wall and have your child draw pictures on this sheet. Drawing on a vertical surface requires more effort and pens train faster;
3. As soon as your child is already quite strong enough to hold small things in his hands and does not let go of them, if you pull slightly, start teaching him how to tie shoelaces or weave pigtails from any ribbons or strings;
4. If you read newspapers or magazines, give your child a marker and ask him to circle all the headlines;
5. A good grip between thumb and forefinger is easiest to develop by shifting beans or even peas from one bowl to another, using only two fingers rather than the entire palm.
6. Frosty windows or foggy bathroom mirrors are a great place to learn to draw with your index finger.

If you wish, you can use each of the ways to develop your child's fine motor skills in everyday life, this will help him learn to write faster in the future.

Institute of Electronic and Information Systems NovSU, [email protected]

The methods of contour analysis, which are optimally used in real-time systems for extracting the contours of objects in a video sequence, are considered.

Keywords: contour, image processing, contour analysis, video surveillance system

Introduction

Image segmentation based on contouring is considered for solving this class of problems due to the fact that changing the parameters of position, rotation and scale of the image has little effect on the amount of computation. In addition, the contours completely determine the shape of the image, are weakly dependent on color and brightness, and contain the necessary information for further classification of the object. This approach makes it possible not to consider the internal points of the image and thereby significantly reduce the amount of processed information due to the transition from the analysis of a function of two variables to a function of one variable. The consequence of this is the possibility of ensuring the operation of the processing system in a time scale closer to real.

Basic concepts

By the image contour we mean a spatially extended break, drop or abrupt change in brightness values.

The ideal drop has the properties of the model shown in Fig. 1a - it is a set of connected pixels, each of which is located next to a rectangular jump in brightness, as shown by the horizontal profile in Fig. In reality, optical limitations, sampling, etc. lead to blurry differences in brightness. As a result, they are more accurately modeled by an inclined profile similar to that shown in Figure 1b. In such a model, the brightness difference point is any point lying on the inclined section of the profile, and the difference itself is a connected set formed by all such points.

Figure 1 Model of ideal (a) and oblique (b) brightness differences

A difference in brightness is considered a contour if its height and tilt angle exceed some threshold values.

Let's note a number of problems that arise during the selection of the contour:

Contour breaks in places where brightness does not change quickly enough;

False contours, due to the presence of noise in the image;

Excessively wide contour lines due to blurring, noise, or due to the shortcomings of the used algorithm;

Inaccurate positioning due to the fact that the contours of the line are one, rather than zero, width.

Differential methods

One of the most obvious and simplest ways to detect boundaries is the differentiation of brightness, considered as a function of spatial coordinates.

The detection of contours for an image with brightness values ​​f (x1, x2), perpendicular to the x1 axis, provides the partial derivative df / dx1, and perpendicular to the x2 axis - the partial derivative df / dx2. These derivatives characterize the rate of change in brightness in the x1 and x2 directions, respectively. To calculate the derivative in an arbitrary direction, you can use the brightness gradient:

grad f (x1, x2) = f (x1, x2).

Gradient is a vector in two-dimensional space, oriented in the direction of the fastest increase of the function f (x1, x2) and having a length proportional to this maximum speed. The modulus of the gradient is calculated by the formula

Figure 2 Graphical representation of the gradient

To select a contour of an arbitrary direction, we will use the brightness field gradients module. For images, instead of derivatives, we take discrete differences.

Roberts operator

One of the options for calculating the discrete gradient is the Roberts operator. Since the difference in any two mutually perpendicular directions can be used to calculate the modulus of the gradient, the diagonal differences are taken in the Roberts operator:

The determination of the difference is formed by two filters with a finite impulse response (FIR filters), the impulse responses of which correspond to 2x2 masks

The disadvantages of this operator include high sensitivity to noise and orientation of region boundaries, the possibility of breaks in the contour, and the absence of a pronounced centering element. And he has one advantage - low resource intensity.

Operators Sobel and Prewitt

In practice, it is more convenient to use the Sobel and Prewitt operators to compute discrete gradients. For the Sobel operator, the effect of corner element noise is slightly less than for the Prewitt operator, which is significant when working with derivatives. Each of the masks has the sum of the coefficients equal to zero, i.e. these operators will give zero response in constant brightness regions.

FIR filters are 3x3 masks.

Sobel operator masks:

Operator Prewitt masks:

The Sobel operator uses a weighting factor of 2 for the middle items. This increased value is used to reduce the anti-aliasing effect by giving more weight to the midpoints.

To address the issue of rotation invariance, so-called diagonal masks are used to detect discontinuities in diagonal directions.

Diagonal masks of the Sobel operator:

Diagonal masks of the Prewitt operator:

In the presence of a central element and low resource intensity, this operator is characterized by a high sensitivity to noise and orientation of the boundaries of areas, as well as the possibility of discontinuities in the contour.

Figure 3. Selection of boundaries by the Sobel operator: a) the original image; b) the result of using the Sobel operator

Laplacian

Differential operators of a higher order can be used to solve the problem of highlighting the differences in brightness, for example, the Laplace operator:

In the discrete case, the Laplace operator can be implemented as a procedure for linear image processing with a 3x3 window. The second derivatives can be approximated by the second differences:

The Laplacian accepts both positive and negative values, therefore, in the operator of the selection of contours, it is necessary to take its absolute value. Thus, we obtain a procedure for selecting boundaries that is insensitive to their orientation

The role of the Laplacian in segmentation problems is reduced to using its zero-crossing property to localize the contour and find out if the pixel in question is on the dark or light side of the contour.

The main disadvantage of the Laplacian is its very high sensitivity to noise. In addition, gaps in the contour are possible, as well as their doubling. Its advantages include the fact that it is insensitive to the orientation of the boundaries of areas, and low resource intensity.

Local processing

Ideally, edge detection methods should only select pixels in the image that lie on the path. In practice, this multitude of pixels rarely renders the outline accurately enough due to noise, edge discontinuities due to irregular lighting, and the like. Therefore, edge detection algorithms are usually complemented by linking routines to generate sets of edge points containing the edges.

One way to link contour points is to analyze the characteristics of pixels in a small neighborhood of each image point that has been marked as contour. All points that are similar according to some criteria are linked and form a contour of pixels that meet these criteria. In this case, two main parameters are used to establish the similarity of the contour pixels: the response value of the gradient operator, which determines the value of the contour pixels, and the direction of the gradient vector.

A pixel in a given neighborhood is combined with a central pixel (x, y) if the criteria for similarity in both size and direction are met. This process is repeated at each point of the image while storing the found associated pixels as the center of the neighborhood moves. A simple way of accounting for data is to assign a different luminance value to each set of contour pixels that are linked.

Canny Border Detector

The Canny boundary detector focuses on three main criteria: good detection (increased signal-to-noise ratio); good localization (correct determination of the position of the border); the only response to one boundary.

From these criteria, an objective function of the cost of errors is constructed, by minimizing which the optimal linear operator for convolution with an image is found.

To reduce the sensitivity of the algorithm to noise, the first derivative of the Gaussian is applied. After applying the filter, the image becomes slightly blurry. This is what the Gaussian mask looks like:

After calculating the gradient of the smoothed image, only the maximum points of the image gradient are left in the border contour. Information about the direction of the border is used in order to remove points exactly near the border and not break the border itself near the local maxima of the gradient.

The Sobel operator is used to determine the direction of the gradient. The resulting direction values ​​are rounded to one of four angles - 0, 45, 90 and 135 degrees.

Then, using two thresholds, weak boundaries are removed. In this case, the fragment of the border is processed as a whole. If the value of the gradient somewhere on the tracked fragment exceeds the upper threshold, then this fragment also remains a "valid" border in those places where the value of the gradient falls below this threshold, until it falls below the lower threshold. If on the whole fragment there is not a single point with a value higher than the upper threshold, then it is deleted. This hysteresis reduces the number of breaks in the output boundaries.

The inclusion of noise reduction in the algorithm increases the robustness of the results, but increases the computational cost and leads to distortion and loss of border detail. The algorithm rounds the corners of objects and destroys the boundaries at the connection points.

The disadvantages of this method are the complexity of implementation and very high resource intensity, as well as the fact that some rounding of the corners of the object is possible, which leads to a change in the parameters of the contour.

The advantages of the method include low sensitivity to noise and the orientation of the boundaries of areas, the fact that it clearly identifies the contour and allows you to identify the internal contours of the object. In addition, it excludes erroneous detection of the contour where there are no objects.

Figure 4. Selection of borders using the Canny method: a) the original image; b) after processing by the Canny algorithm

Analysis with graph theory

Based on a graph representation and a search on that graph for least-cost paths that correspond to significant contours, you can build a method that works well in the presence of noise. This procedure turns out to be rather complicated and requires more processing time.

Figure 5. A contour element located between pixels p and q

A contour element is a border between two pixels p and q, which are neighbors. The contour elements are identified by the coordinates of the points p and q. The contour element in Fig. 5 is defined by pairs (xp, ur) (xq, yq). A contour is a sequence of contour elements connected to each other.

The problem of finding the minimum cost path on the graph is non-trivial in terms of computational complexity, and one has to sacrifice optimality in favor of the computational speed.

The complexity of implementation and high resource consumption are the main disadvantages of such an analysis, the advantage of which is its low sensitivity to noise.

Conclusion

The methods presented in the work describe the optimal approaches for extracting contours in real-time systems. The methods allow solving a wide range of contouring tasks, which are used in many areas where image segmentation is necessary.

Literature

1. Gonzalez R., Woods R. Digital image processing. M .: Tekhnosfera, 2005.S. 812-850.

2. Yane B. Digital image processing. M .: Technosphere, 2007. С.331-356.

3. Methods of Computer Image Processing / Ed. V.A. Soifer. Moscow: Fizmatlit, 2003.S. 192-203.

4. Pret U. Digital image processing. M .: Mir, 1982.S. 499-512.

5. See: http://www.cs.berkeley.edu/~jfc/