以基于样例修补的目标移除方法为基础,改进了基于样本块的图像纹理修补方法.首先将抠像技术应用到目标物的提取过程中,然后使用图像分割的方法实现了分区修补的目的.改进后的方法能够更好地还原图像的线性特征并且减少图像修补的时间.在此基础上,根据不同角度的图像可以提供更加丰富的修补信息这一思想,进一步提出一种基于多幅图像的修补方法,并取得了更高质量的修补效果.实验结果表明,该方法不仅能够准确地提取并移除图像上不需要的物体,而且能够有效地修补图像受损区域的纹理和结构.
An improved example-based inpainting method of image texture is proposed in the paper. By the method, first an image matting technique is used to extract the object, and then an image segmentation method is used to realize inpainting in sub-areas. The improved method can maintain the structure of the image and reduce the processing time. Further more, by taking advantage of richer information of multiple images from different point of views, a multiple images-based inpainting method is proposed to obtain the result in higher quantity. Experimental results show that the improved method can remove the object precisely and inpaint the structure and texture effectively.