图像收缩是缩小高分辨率图像以适应不同纵横比小尺寸显示屏幕的过程,关键是收缩后能够凸显图像重要区域,保持连续,避免扭曲.提出一种新的图像收缩方法,该方法首先基于能量失真约束,迭代收缩覆盖图像的四边形网格至目标大小,然后映射,插值目标网格实现图像收缩.能量失真反映了对重要区域的凸显程度、结构的保持效果以及扭曲避免情况,失真越小,目标图像越理想.在该约束下,构成网格的子四边形非均匀收缩,重要度大的收缩小.为准确计算子四边形的重要度,根据图像显著度和边缘构建反映图像重要度的Hot-Target图.最后,通过保持图像直线边,称为特征边缘,避免非均匀收缩引起的边缘扭曲.为提高效率,降低复杂度,该方法由迭代求解线性方程实现.实验结果验证了方法的有效性.%Image shrinkage is the process of reducing image resolution to adapt to display screens with different aspect ratios and different sizes. Its key is to highlight important areas, keep continuity and avoid twists. This paper presents a novel image shrinking method. First, this paper iteratively shrinks the quad mesh covering the original image to the target size under the constraint of energy distortion. Then, this paper obtains the arget image by interpolating and mapping the target mesh. The energy distortion function reflects the effects of highlighting important regions, preserving structure and avoiding twists. Less distortion owns better result. Under the constraint of energy distortion, every sub quad of mesh shrinks non-uniformly. Quads with more importance shrink less. In order to accurately calculate quad's importance, this paper proposes a new method named as Hot-Target map to calculate image importance according to image saliency and edges. Finally, this paper avoids distortion by preserving image linear edges named as featured straight edge. To increase efficiency and reduce complexity, the method is carried out by solving linear equations. Experimental results verify its effectiveness.
展开▼
