Genetic analyses of coat colors are frequently restricted to subjectively categorized phenotype information. The aim of this study was to develop a method to numerically quantify the variability of leopard complex spotting phenotypes introducing tools from image analysis. Generalized Procrustes Analysis eliminates systematic errors due to imaging process. The binarization of normalised images and the application of Principal Component Analysis on the derived pixel matrices, transform pixel information into numerical data space. We applied these methods on 90 images to ascertain the specific leopard patterns within the Noriker breed. Furthermore we genotyped a representative sample of 191 Noriker horses for the known leopard complex spotting associated loci. 97% of the genotyped leopard spotted horses were heterozygous for LP and had at least one copy of the PATN1 allele. However, the remaining pattern variation was great, indicating other genetic factors influencing the expression of leopard complex spotting. Based upon this data we estimated effect sizes of the modifier PATN1, and additional factors including sex, age, base color and spotting phenotype of parents. The Principal Component Analysis of the pixel matrix resulted in two significant components accounting for 51% of the variation. Applying a linear model, we identified significant effects for age groups and base color on the first and second components, whilst for sex and parents" LP phenotype significant effects were found on four additional components.