FIXME **This page is not fully translated, yet. Please help completing the translation.**\\ 
// (remove this paragraph once the translation is finished) //

==== MTF/SFR Measurement — Dead Leaves Method ====
**Principle:**\\
Based on a random circular texture pattern that closely resembles natural scenes, this method extracts the texture Power Spectral Density (PSD) and subtracts the noise. It then calculates the ratio of the modulation amplitude at each frequency to the target modulation to obtain the MTF/SFR values. It is a resolution test used for specific scenarios, namely texture loss (such as high-frequency details under low-contrast conditions, e.g., hair).
{{http://ydadmin.rdbuy.com.cn/ueditor/php/upload/image/20230814/1691996772313271.png}}
**Advantages:**\\
The statistical properties of the pattern closely match those of real images, and the measurement results align well with the human eye's subjective perception of texture sharpness.\\
It is sensitive to image noise reduction algorithms and can effectively evaluate the loss of fine textures caused by noise reduction.\\

**Disadvantages:**\\
It requires a relatively large image area, resulting in lower spatial utilization efficiency compared to the slanted-edge method.\\
The computational process is complex and relies on precise noise correction and texture extraction algorithms.\\

**Test Chart:**\\
The Dead Leaves chart should comply with the CPIQ/IEEE 1858 standard (contrast 3:1) and consists of randomly stacked circular patterns; the selection should match the specific testing requirements. The color Dead Leaves chart can accurately capture both chrominance and luminance texture loss, yielding results that closely match subjective perception; the monochrome Dead Leaves chart detects only luminance texture loss and provides more stable measurements under low-light or high-noise conditions.\\
Color Dead Leaves chart:\\
|{{ http://ydadmin.rdbuy.com.cn/ueditor/php/upload/image/20240408/1712555414222957.png}}|

Monochrome Dead Leaves chart:\\
|{{ http://ydadmin.rdbuy.com.cn/ueditor/php/upload/image/20240408/1712555440145864.png}}|
        

**See More**\\
[[yanding:成像质量评价:测试用例:camera测试用例:纹理丢失]]