RGB＆sRGB Colorchecker数据的多元多项式回归

Question

我需要将RGB图像映射到ColorChecker的sRGB，并使用这样获得的系数对数据库中的其他（舌图像）图像进行颜色校正

我尝试过线性回归，但不确定在多元线性回归的情况下是否正确，文献建议多项式回归是更好的预测

df=pd.read_csv("colorchart_photo_strobo_linear.csv")
df=pd.read_csv("colorchart_photo_strobo_linear.csv")

    y=df.iloc[:,1:4].values 
    yr=df.iloc[:,1:2].values
    yg=df.iloc[:,2:3].values
    yb=df.iloc[:,3:4].values 
    df1=pd.read_csv("colorchart_rendered_strobo_linear.csv")
    x=df.iloc[:,1:4].values
    plt.scatter(x, y, color='red')
    plt.grid(True)

     regr1 = linear_model.LinearRegression()
     regr.fit(x, yr)

     print('Intercept: \n', regr.intercept_)
     print('Coefficients: \n', regr.coef_)
     plt.plot(x,y,color='blue')
     plt.show()
     regr2 = linear_model.LinearRegression()
     regr.fit(x, yg)

     print('Intercept: \n', regr.intercept_)
     print('Coefficients: \n', regr.coef_)
     plt.plot(x,y,color='green')
     plt.show()
     regr3 = linear_model.LinearRegression()
     regr.fit(x, yb)

     print('Intercept: \n', regr.intercept_)
     print('Coefficients: \n', regr.coef_)
     plt.plot(x,y,color='magenta')
     plt.show()
     #prediction with sklearn
     x = [[0.14703, 0.10224,    0.0319]]
     print ('printyr: \n', regr.predict(x))

需要的是Xout = [A] Transpose.Xin