Research on multi feature color extraction method of network illustration based on feature matching

In order to improve the accuracy and efficiency of multi feature color extraction of network illustration, this paper proposes a new method of multi feature color extraction of network illustration based on feature matching. According to the characteristic value and the change track of the pixels of the illustration image, the color characteristic value of the illustration image is extracted by Gaussian filter, and the Euclidean distance is used to complete the matching calculation. Based on the results of feature matching, the color of illustration image is transformed from RGB space to HSV space, and the multi feature color extraction of network illustration is completed according to the results of matrix dimensionality reduction. Experimental results show that, compared with the traditional color extraction method, the proposed color extraction method has higher extraction accuracy and efficiency. It shows that the multi feature color extraction method of this method can improve the extraction performance and practical application value.