Using Cohen-Or method to harmonize a color-rich image leads to severe alterations in the original colors of an image and thus creates a processed image that is different from the original one. Furthermore

deciding on the type of optimal color harmonization template is time consuming. This study introduces skewness and kurtosis

which are used to describe all variant values in the form of deviation and steep degree

to calculate the scope of the value H of a harmonization template gray area and categorize seven harmonization templates into X and T types. According to the gray area loss of at least the principle of the optimal template

the result of value H is applied to the gray area in the X or T harmonization template. Normalized Gaussian function is commonly used to migrate color into the gray area of the adjusted color harmonization template. Harmonization of the original image is then achieved. Experimental results show that the harmonized image using the proposed method exhibits great similarity to the original one

retains the original colors

and considerably reduces the time needed in the process of harmonization. The harmonious processing time of a 450×423 image is reduced from 2 549.78 s to 13.869 3 s. The harmonized image using the proposed method retains its color style similarity to the original image. The proposed method facilitates grayscale image colorization

image color transfer

and industrial product style transfer.