In this project, we reproduce the “Split and Match” method proposed by O. Frigo et al.[1] for unsupervised style transfer. The method interprets image style as a combination of local texture and global color, transferring style from an example image to a content image while preserving its structural content. The key of this approach is an adaptive partition that can align image patches based on similarity with example image, enabling effective local texture transfer. We implement and test the method on various image pairs to verify its performance, with a focus on texture than color. Our reproduction confirms the effectiveness of the proposed technique and offers insights into its strengths and limitations under different conditions.