Landscape Information Sketching Integrating Image Structural Features

This study presents an innovative method for the landscape information sketching that integrates image structural features to enhance the artistic quality and expressiveness of the resulting sketch images. Traditional approaches rely heavily on spatial gradient processing, which can lead to computational complexity and unsatisfactory results. To mitigate these challenges, this study proposes a model that utilizes image structural features, specifically an enhanced version of the Scale-Invariant Feature Transform (SIFT), to extract critical information such as object contours and spatial distribution. The proposed sketching algorithm uses a least-squares method to linearly represent image data, facilitating the accurate synthesis of sketched images from existing image libraries. Experimental results illustrate the effectiveness of our approach in producing sketch images with varying levels of detail and style, providing digital artists and designers with flexibility and control. In addition, quantitative evaluation metrics such as SSIM, PSNR, and MSSIM highlight the potential for the further optimization and improvement, suggesting directions for future research in advanced image processing techniques and the integration of deep learning models.