| Summary: | 碩士 === 國立成功大學 === 資訊工程學系 === 104 === We propose a novel method to generate calligraphy art from a single input image by replacing the partial curves of the image with the best matched shape of characters. The salient structure of the image is preserved, and the contour is reconstructed with characters. Most previous word arts focus on the word filling process inside a specific contour. In contrast to word cloud, our work focuses on the replacement process to find best matching character to the structure of the input image. In our framework, we first segment the input image into regions, and convert a character into a single curve. Second, We analyze the structure of the region contour and the curve of character for finding the relationship between the structure of the character and the salient contour in the image. We proposed a modified partial curve matching to generate a stylized word art composite with character. A series of comparative results are generated to demonstrate the flexibility of the constraint metrics, and the results of the proposed system are visually compelling.
|
|---|
