First, the principle of digital image clarity
Sharpness is the sharpness of the edge of the image detail. At the edge of the image detail, the sharper (faster change) or sharper (high contrast) the optical density or brightness changes with position, the clearer the edge of the detail and the higher the visibility.
During the transmission of image information, the details themselves and their clarity will be lost. Taking print copying as an example, assuming that the original is a color reversal film, its detail resolution (resolution) can reach or even exceed 100 line pairs/mm, but according to the sampling theorem, only the scanning resolution of more than 200 line pairs/mm (5080ppi). In order to distinguish all the details. It can be seen that more image detail information is lost under the commonly used scanning resolution. The rose spots formed after the color image is screened also interfere with the reproduction of the details, making it difficult to discern. Factors that affect the clarity of printed images are image contrast compression, plate-making and printing register accuracy. In order to ensure the reproduction of the image details and its sharpness, it is necessary to perform image sharpness enhancement processing so that details that have not been lost are well reproduced.
The details of the image tend to have more obvious boundary optical density jumps. The human-eye vision system has the feature of enhancing the boundary area where the image density is abrupt. The old photo-engraving technology uses Unsharp Mask technology to increase the contrast of the density transition edge to achieve the effect of accentuating sharpness.
In today's digital image acquisition and processing, the resolution enhancement is also processed in the form of digitization. The original image is digitized into discrete pixels after scanning, and the pixels have different grayscale values. At the edge of the density of the image details, there are also significant differences in the gray value of the digital image signal.
Digital virtual light mask technology is widely used in image scanners and image processing software. The basic principle is that when processing a certain pixel, several pixels around the pixel are taken, and the mean U average of the gray levels of these surrounding pixels is calculated. Then, the gray average value U average of the surrounding pixels is subtracted from the gray value U center of the center pixel to obtain the virtual mask signal U virtual light mask. Adding this virtual mask signal to the gray value of the center pixel allows the contrast of the edge of the image to change, that is, the dark side of the edge is darker and the brighter side is brighter. The contrast of the edge of the enlarged detail can be used to improve the visual clarity.
For different image sharpness processing requirements, different intensity intensities of ghost mask signals can be processed (amplitude value k), and the range of the surrounding pixels (radius value) can also be changed. At the same time, in order to avoid rough feeling caused by the sharpness of the skin and other regions, a threshold U threshold can be set. Only when the contrast of the details of the edge gray value is greater than the threshold, the definition is emphasized. Some excellent software can even emphasize the sharpness of different gradations and different hue colors.
Second, sharpness correction adjustment methods and points
Take Photoshop as an example, the Unsharp Mask function has three adjustable items, namely the amplitude, radius, and threshold. The magnitude of the amplitude value directly determines the strength of the sharpness emphasis; the size of the radius value determines the border width involved in the sharpness emphasis. If the radius value is large, the detail contour sense is enhanced; the threshold value causes the contrast smaller detail to be affected. Protection is not emphasized, the higher the threshold, the more detail is protected. The correct use of the above three adjustment items helps the correct reproduction of image clarity.
In the image sharpness adjustment processing, four basic points should be noted.
1. Take different settings for different body type manuscripts
Images that typically have large areas of skin should be emphasized with a lower degree of sharpness, or higher thresholds used to protect skin detail from being too rough.
2. Different settings for images with different formats and resolutions are used. Sharpness enhancement is performed after the image is calibrated, the output size, and the resolution have been set. In general, for images with low resolution and small size (such as 72ppi web images), smaller radius values ​​should be used to avoid strong contour relief. For images with high resolution and large size, slightly larger radius values ​​can be used.
3. The definition of clear manuscript emphasizes the principle
If the original image itself is clear, the emphasis is 100% to 150%. The adjustment of image clarity as a whole should grasp the "clear and mellow" dimension, abandoning the rough and clear effects.
4. Defining the principle of the original resolution of printed matter
Some printed manuscripts that have gone through the net, when emphasizing sharpness, should pay attention to the adjustment amount not to be too large, otherwise prone to moire.
Third, the clarity of the channel adjustment method
In the definition of color image adjustment, it is usually a plurality of primary color channels to adjust together. This adjustment method sometimes leads to an abnormal phenomenon of edge dew, especially at the border between color and white.
Take the red/white junction as an example: In the CMYK mode, there are obvious boundary brightness jumps on the yellow channel and the magenta channel, and the blue channel has very weak luminance jump at the red/white junction. After the imaginary mask is made, the yellow/magenta channel produces a “blacker/whiter†boundary jump at the border, while the blue channel changes less. Some of the blue will be overprinted in the “whiter†position of the yellow/magenta border, exposing the cyan border.
The sharpness adjustment of the color image can be divided into channels, so as to avoid the detail edge dew in the image.
For CMYK mode color images, according to their main colors, different color channels can be emphasized for clarity.
It can be seen that for the warm-color image, when the clarity of all its channels or the yellow/magenta channel is emphasized, there will be a dew phenomenon at the edge of the details of the image. When only the black channel or the blue/black channel is emphasized, the delustering phenomenon is less obvious.
For L*a*b* mode color images, only the L (brightness) channel can be sharpened to avoid edge dew.
It can be seen that the CMYK and L*a*b* mode color image resolution channel adjustment methods increase the sharpness adjustment without adverse effects and have a good effect.
Adjustment and correction of image sharpness is a delicate task. As long as the images are adjusted according to the different characteristics and types of the images, the functions can be properly adjusted based on the balanced use of the functions. Thus, satisfactory results can be obtained.
Source: "Printing Technology"