Sepia¶

Applies sepia effect to image.

int sepia(cv::InputArray src, cv::OutputArray dst)¶

Parameters:	src – Source 8-bit single-channel image. dst – Destination image of the same size and the same type as src.
Returns:	Error code.

The algorithm.

1. Create 3-channel image, which is interpreted as HSV image.

   1. 1st channel is the matrix, each element equals hue = 19.
   2. 2nd channel is the matrix, each element equals saturation = 78.
   3. 3rd channel is the matrix of brightness (src + value = src + 20).

2. Convert resulting image to BGR format.

Example.

Antique¶

Applies antique effect to image.

int antique(cv::InputArray src, cv::InputArray texture, cv::OutputArray dst)¶

Parameters:	src – RGB image. texture – RGB image that is overlaid with the src image as a texture of an old photo. dst – Destination image of the same size and the same type as src.
Returns:	Error code.

The algorithm.

1. Resize texture to match the source image size.
2. Apply sepia matrix transform to the src.
3. Calculate the weighted sum of the texture and src.
4. Convert resulting image to the same color format as src.

Example.

Vignette¶

Makes the edges of the photo less bright, creating an oval frame around its central part.

int vignette(cv::InputArray src, cv::OutputArray dst, cv::Size rect)¶

Parameters:	src – Source 8-bit three-channel image. dst – Destination image of the same size and the same type as src. rect – Size of rectangle describes an ellipse, whose center is at the center image.
Returns:	Error code.

The algorithm.

1. Create a new 3-channel image, which is interpreted as BGR image.

2. For every channel calculate \(dist\), \(radius\_ellipse\) and \(coefficient\), where \(dist\) is a distance between the current pixel and the center of image; \(radius\_ellipse\) is a radius of ellipse in the this point which belongs the same line as a pixel; \(coefficient\) is a number which multiplies the intensity channel.

   The \(coefficient\) is calculated by the following formula:

   \[coefficient = 1 - ((dist - radius\_ellipse) / (max\_radius - radius\_ellipse)),\]

   where \(max\_radius\) is a distance between the pixel (0, img_src.cols) and the center image.

3. Convert image to a BGR format.

Example.

Matte¶

Increases the brightness of peripheral pixels.

int matte(cv::InputArray src, cv::OutputArray dst, cv::Point firstPoint, cv::Point secondPoint, float sigma)¶

Parameters:	src – RGB image. dst – Destination image of the same size and the same type as src. firstPoint – The first point for creating ellipse. secondPoint – The second point for creatin ellipse. sigma – The deviation in the Gaussian blur effect.
Returns:	Error code.

The algorithm.

1. Create new image with white background for mask.
2. Draw black ellipse inscribed in a rectangle that is defined by two opposite corner points (firstPoint and secondPoint) on the mask image. It’s a meaning part.
3. Apply gaussian blur to the meaning part to make fade effect.
4. Convolve mask with the image.
5. Convert resulting image to the same color format as src.

Example.

sigma = 25.

Edge Blur¶

Blurs the edges of an image, keeping center in focus.

int edgeBlur(InputArray src, OutputArray dst, int indentTop, int indentLeft)¶

Parameters:	src – RGB image. dst – Destination image of the same size and the same type as src. indentTop – The indent from the top and the bottom of the image. It will be blured. indentLeft – The indent from the left and right side of the image. It will be blured.
Returns:	Error code.

The algorithm.

The amount of blurring is defined by the gaussian filter’s kernel size and standard deviation and depends on the weighted distance from the center of the image:

\[d(x, y) = \frac{(x - a)^2}{(a - indentLeft)^2} + \frac{(y - b)^2}{(b - indentTop)^2},\]

where \(a = \frac{src_{width}}{2}, b = \frac{src_{height}}{2}\). For each pixel \((x, y)\) of the image, if the distance \(d(x, y)\) is greater than 1, gaussian filter with center at \((x,y)\), kernel size \(d(x, y)\) and standard deviation \((radius - 0.5)\) is applied, otherwise src image pixel is left unchanged. Border pixels are replicated to fit the kernel size.

Example.

indentTop = 90, indentLeft = 90.

Fade Color¶

Applies color fade effect to image.

int fadeColor(InputArray src, OutputArray dst, Point startPoint, Point endPoint)¶

Parameters:	src – Grayscale or RGB image. dst – Destination image of the same size and the same type as src. startPoint – Initial point of direction vector for color fading. endPoint – Terminal point of direction vector for color fading.
Returns:	Error code.

The algorithm.

1. Determine the coordinates of the vector by two points (startPoint, endPoint) .

2. Determine the line which is perpendicular to vector and is passing through startPoint.

3. Find the most distant point from the line.

4. For each pixel located at one side from the line defined by the direction of the vector, change the value of each channel by the following formula:

   newValue = (1-a) * oldValue + a * 255, a = distance / maxDistance.

5. Save this matrix as image in same format.

Example.

Boost Color¶

Makes colors more saturated.

int boostColor(cv::InputArray src, cv::OutputArray dst, float intensity=0.0f)¶

Parameters:	src – RGB image. dst – Destination image of the same size and the same type as src. intensity – Effect intensity, must be real number from 0.0 to 1.0.
Returns:	Error code.

The algorithm.

1. Create 3-channel image, which is interpreted as HLS image
2. Each element of the 3rd (Saturation) channel increases by intensity
3. Save this matrix as RGB image

Example.

Tint¶

Add a tint to the image.

int tint(cv::InputArray src, cv::OutputArray dst, const cv::Vec3b& colorTint, float density)¶

Parameters:	src – Source 8-bit 3-channel (RGB) image. dst – Destination image of the same size and the same type as src. colorTint – It’s a bearing color. All color of the image src will be shifted to it. density – Float value between 0 and 1, defines a range of shift to the colorTint.
Returns:	Error code.

The algorithm:

Calculate new values by the formula:

\[dst(x, y) = density \cdot src(x, y) + (1 - density) \cdot colorTint\]

Example:

density = 0.1, colorTint = Vec3b(255, 255, 0), i.e. cyan.

Warmify¶

Increases saturation of red and yellow tones, making photographs more warm, sunset view.

int warmify(cv::InputArray src, cv::OutputArray dst, uchar delta=30)¶

Parameters:	src – Source 8-bit three-channel image. dst – Destination image of the same size and the same type as src. delta – Value by which saturation of warm colors is increased.
Returns:	Error code.

The algorithm.

1. Create 3-channel image, which is interpreted as BGR image.

   1. 1st channel is the matrix, each element equals blue = blue_src.
   2. 2nd channel is the matrix, each element equals green = green_src + delta.
   3. 3rd channel is the matrix, each element equals red = red_src + delta.

2. Save this matrix as BGR image.

Example.

Glow¶

Applies glow effect to the initial image

int glow(cv::InputArray src, cv::OutputArray dst, int radius=0, float intensity=0.0f)¶

Parameters:	src – RGB image. dst – Destination image of the same size and the same type as src. radius – Radius of box filter kernel, must be positive integer number intensity – Effect intensity, must be real number from 0.0 to 1.0
Returns:	Error code.

The algorithm.

1. Create the copy of the source image

2. Apply box filter for this copy with deviation equal sigma

3. Create new 3-channel image, each channel of the matrix calculates by the following formula:

   \[\begin{split}C = \begin{cases} 2 * A * B, & \text{if $B <= 0.5$}\\ 1 - 2 * (1 - A) * (1 - B), & \text{otherwise} \end{cases}\end{split}\]

   where \(A\) is the pixel’s tonal value that lies in the blurred image, \(B\) is the pixel’s tonal value that lies in the source image, \(C\) is the tonal value of the blended pixel.

4. Create new 3-channel image, each channel of the matrix calculates by the following formula:

   \[C = intensity * A + (1 - intensity) * B\]

   where \(A\) is the pixel’s tonal value that lies in the previous blended image, \(B\) is the pixel’s tonal value that lies in the source image, \(C\) is the tonal value of the blended pixel.

5. Save previous image as RGB image

Example.

Film Grain¶

Applies film grain effect to the initial image.

int filmGrain(InputArray src, OutputArray dst, int grainValue, RNG& rng)¶

Parameters:	src – Grayscale or RGB image. dst – Destination image of the same size and the same type as src. grainValue – Degree of graininess. 8 is default value. rng – Random number generator. cv::theRNG() is default value.
Returns:	Error code.

The algorithm.

1. Create matrix with noise.
2. Add noise to image.

Example.