opencv使用 --- fastGlobalSmootherFilter

0. 背景

做人脸数据集处理的时候，需要做光照合成，在 Face Illumination Transfer through Edge-preserving Filters [1] 里介绍了一种光照迁移方法，需要用到 WSL [2] edge-preserving Filters。

1. opencv 接口

opencv有WSL原理的改进版FGS[3]， 具体接口[4]形式：

void cv::ximgproc::fastGlobalSmootherFilter     (     InputArray      guide,
        InputArray      src,
        OutputArray      dst,
        double      lambda,
        double      sigma_color,
        double      lambda_attenuation = 0.25,
        int      num_iter = 3 
    )         

# Python:
     dst    =    cv.ximgproc.fastGlobalSmootherFilter(    guide, src, lambda, sigma_color[, dst[, lambda_attenuation[, num_iter]]]    )

2. 使用方法

需要先配置 opencv-contrib包，C++的配置没有具体尝试，python需要 pip install opencv-contrib-python 才能 import cv2.ximgproc

C++使用参考源码的 fgs_test

TEST(FastGlobalSmootherTest, SplatSurfaceAccuracy)
{
    RNG rnd(0);

    for (int i = 0; i < 5; i++)
    {
        Size sz(rnd.uniform(512, 1024), rnd.uniform(512, 1024));

        int guideCn = rnd.uniform(1, 2);
        if(guideCn==2) guideCn++; //1 or 3 channels
        Mat guide(sz, CV_MAKE_TYPE(CV_8U, guideCn));
        randu(guide, 0, 255);

        Scalar surfaceValue;
        int srcCn = rnd.uniform(1, 4);
        rnd.fill(surfaceValue, RNG::UNIFORM, 0, 255);
        Mat src(sz, CV_MAKE_TYPE(CV_16S, srcCn), surfaceValue);

        double lambda = rnd.uniform(100, 10000);
        double sigma  = rnd.uniform(1.0, 100.0);

        Mat res;
        fastGlobalSmootherFilter(guide, src, res, lambda, sigma);

        // When filtering a constant image we should get the same image:
        double normL1 = cvtest::norm(src, res, NORM_L1)/src.total()/src.channels();
        EXPECT_LE(normL1, 1.0/64);
    }
}

python 方法

import cv2
from cv2.ximgproc import *
img = cv2.imread('test.jpg')
imgSmooth = fastGlobalSmootherFilter(img, img, 625.0, 20)

注： guide 可以用原图或者灰度图，其他类型没有尝试

        在python中 lambda是关键字，所以不能用 lambda=625.0这样的传参方式

        sigma_color 参数取值 1.0~100， 低于1.0 看不出效果（跟论文里的参数设置情况不太一样）

参考：

[1] X. Chen, M. Chen, X. Jin, and Q. Zhao, “Face illumination transfer through edge-preserving filters,” in Proc. CVPR, 2011, pp. 281–287.

[2] Z. Farbman, R. Fattal, D. Lischinski, and R. Szeliski, “Edge-preserving decompositions for multi-scale tone and detail manipulation,” ACM Transactions on Graphics (Proc. SIGGRAPH), vol. 27, no. 3, Aug. 2008.

[3] Dongbo Min, Sunghwan Choi, Jiangbo Lu, Bumsub Ham, Kwanghoon Sohn, and Minh N Do. Fast global image smoothing based on weighted least squares. Image Processing, IEEE Transactions on, 23(12):5638–5653, 2014.

[4] OpenCV-3.4.4