How to save the image quality--加拿大Paper代写范文

加拿大Paper代写范文:“How to save the image quality”，这篇论文主要描述的是如何保存图像的质量，我们可以通过补偿背光和调光的方法来实现，这样能够使得一些精美的画面质量不会大打折扣，减少伽马的值，照亮在暗区中的图像，在加上升压剪辑和通道算法的优化，能够让我们保留住图像画面的高光细节。

We described two methods for preserving image quality while compensating for backlight dimming.  These give a graceful reduction in image quality as the image backlight is reduced below the brightness of the image.

The algorithms are non-adaptive and hence void of temporal artifacts which can afflict image adaptive processes such as independent frame histogram equalization.

We have primarily used this method for compensating for a fixed BL reduction but it may be applied as part of an adaptive BL system, allowing more aggressive BL dimming.  The lack of temporal artifacts which may occur with image adaptive algorithms such as histogram equalization is a definite advantage. 

Comparison with reducing gamma

A well known technique for brightening an image is to process it to reduce gamma.  Reducing gamma is a point wise tone-scale operation and can be compared with our simpler algorithm using the proposed BP tone-scale.  When the approach using gamma reduction is evaluated in the brightness compensation application, it is found that a reduction in gamma only gives accurate compensation at a single code value (e.g., 100), in the Figure 5.  Code values above this are not sufficiently brightened performing worse than the proposed method.  

Figure 5 Comparison with reducing gamma

As seen in Figure 6, reducing gamma over- brightens the dark region of the image resulting in a washed out look.  Further reduction in gamma to increase the brightening at high code values only increases the black level further.

Image quality comparison

We compare the performance of the boost-clip, smooth roll-off, and 2-channel algorithms in preserving highlight detail.  These algorithms all give the same result when for image data below the MFP.  They differ in their handling of highlight detail.  An original image with highlight detail is processed by three algorithms assuming a 20% reduction backlight illumination.

Figure 7 Original Image

The wave contains highlight detail which is above the range of a display with 80% backlight.  Image code values above 230 cannot be represented on the low power display.  We selected an MFP of 180 as appropriate for 80% power.

In Figure 8, we use pseudo color to indicate those pixels in the original which are clipped in magenta, and those in the roll-off region in cyan. 

Figure 8 Distorted pixels in pseudo color

Focusing on the clipped region, in Figure 9 we see the difference in the algorithms.  Clipping gives the greatest loss of detail.  The roll-off attenuates contrast, and hence detail.  The two-channel algorithm gives the greatest contrast preservation.

51due留学教育原创版权郑重声明：原创留学生作业代写范文源自编辑创作，未经官方许可，网站谢绝转载。对于侵权行为，未经同意的情况下，51Due有权追究法律责任。

51due为留学生提供最好的服务，亲们可以进入主页了解和获取更多paper代写范文 提供美国作业代写以及加拿大math辅导服务，详情可以咨询我们的客服QQ:800020041哟。-xz