Combination Of Serial Screen And Artificial Intelligence

Have you ever encountered the situation that the pictures taken by the camera are imported to the computer and displayed differently?

This article will analyze this phenomenon from the matching of Gamma correction and Degamma correction.

Why Gamma Correction

1. From the visual characteristics of the analysis
Physical gray scale, that is, the gray value of the linear growth of the gray scale, the human eye observes the gray scale will think that the bright tones occupy most of the area, the darker tones only account for a small portion of the indistinguishable.

Visual gray scale, that is, the human eye observes the uniform growth of gray scale, the human eye observes that the proportion of bright tones and dark tones is uniform, but the actual gray scale value change is not uniform.

Weber in the visual brightness perception of the experiment, found that the human eye’s perception of brightness is not linear, the perception of dark area tones will be more sensitive than the perception of bright area tones.Here, the physical gray scale and visual gray scale are placed in a twodimensional coordinate system, and it is found that the human eye’s response to luminance under dark tones will be significantly higher than that of bright tones, such as when the natural gray scale is 0.2, and the visual gray scale has already reached 0.5. Therefore, in order to obtain good visual effects, it is necessary to make non-linear changes to the collected linear signals, so as to make them conform to the way of observation of the human eye.

In fact, this principle has been followed throughout the history of camera development. Early film cameras, relying on the nonlinear adaptation of the sensitization of silver salt particles to the human eye perception, through the chemical rocess of the film developed photographs to be presented.

In the era of digital cameras, the earliest use of CRT monitors to display pictures, but the brightness of the CRT monitor and the input voltage was γ ≈ 2.2 non-linear relationship, so it is necessary to Gamma correction on the camera side, to solve the direct input of linear signals will lead to the loss of the dark part of the image, the bright part of the overexposure, which is the origin of the Gamma correction in the ISP of the digital camera.

With the emergence of new technologies LCD, OLED and other displays, although the brightness of the display and the input voltage is basically linear,but in order to be compatible with CRT monitors and related video technology, so on the imitation of CRT monitor characteristics, the display in the reception of the input data, the display will be the first Degamma correction, and has become the general standard of the display industry. 2, from the camera’s characteristics of the analysis of the camera using the Specific precision will be the real world of analog signals into the camera in a
specific precision of the digital signal. Cameras with different precision presents different imaging effects, but does not mean that the higher the precision the more appropriate, the higher the data precision the larger the storage space required, the tighter the transmission bandwidth requirements, the better the image details presented, but the details brought about by the high precision of the enhancement is not necessarily distinguishable by the naked eye.

Weber experiments, found in the medium brightness range only the human eye can distinguish between about 100-200 levels of gray (common RGB888 color space, using 8bit precision provided by 256 levels is also the reason). So observe the following figure more than 7bit grayscale precision, do not look closely to analyze the details usually can not see the difference.

When the camera uses 5bit precision, to refer to the natural gray scale, if you directly use linear changes in the gray scale more than half of the area are highlights, and the differences between the highlights, the human eye is difficult to analyze, if you use the gamma-corrected grayscale, the human eye can be resolved by the gray scale difference is much more.

The introduction of Gamma correction is often inseparable from Degamma correction, Gamma correction, Degamma correction is a pair of complementary nonlinear transformation operations. Both corrections are widely used in image compression and image processing. Gamma correction, also known as positive Gamma, is a power function through the linear signal mapped into a nonlinear signal technology.Degamma correction, also known as inverse Gamma, is a power function through the nonlinear signal mapped into a linear signal technology.

Camera and monitor systems have their own industry-standard grayscale factor settings to ensure that consistent images are displayed on different hardware and software platforms.

At the imaging system end, the raw data of the RAW image captured by the camera responds linearly to light intensity, while the data is also highly accurate (12/14bit). However, the human eye can’t fully distinguish it, so the
non-linear mapping of pixel values can be realized by Gamma correction in ISP, which reduces the image precision under the premise of ensuring the effect observed by the human eye, and realizes the dynamic range of compressed image brightness, optimizes storage space, and reduces the transmission bandwidth.

On the display system side, if it is a CRT display, it comes with a nonlinear relationship of γ ≈ 2.2 itself, while LCD/OLED and other displays will be Degamma-corrected first, calibrated to a specific value of the grayscale coefficient, and the display data will be parsed and displayed.

The process can be shown in this superimposed graph:

By applying gamma correction twice, once during shooting and once during display, it is ensured that what we see on the screen matches the real-life scene. This process compensates for both the limitations of the digital sensor and the characteristics of the display device. With this combination of positive and negative Gamma, the system is able to achieve high-quality image reproduction that matches the perception of the human eye with limited bandwidth and storage.
In addition, Degamma correction is also widely used in ISP algorithms that need to be processed in the linear light domain, such as 3A, de-mosaicing, noise reduction, HDR, etc. For example, in a major ISP, degamma correction is performed before the de-mosaicing algorithm.

Degamma correction before HDR algorithm in a major ISP algorithm.

How to do Gamma correction

Gamma correction adopts the lookup table method, it can work in RGB domain and do gamma correction for R, G and B channels. It can also be converted to YUV domain to do Gamma correction for Y channel only.

Effect of Gamma Correction

Gamma change will change the brightness and contrast of the image, but different from brightness and contrast adjustment, Gamma correction is nonlinear, different pixel values are adjusted differently, Gamma correction retains
shadows and highlights.

NoseDisplay 2025.07.28 News