基于特征加权融合算法的动漫人物图像复原重建方法

A restoration and reconstruction method for anime character images based on feature weighted fusion algorithm

动漫人物复原图像存在像素缺失会导致所设计人物出现失真显示的情况,从而使主机元件无法准确重建动漫人物模型。针对此问题,提出基于特征加权融合算法的动漫人物图像复原重建方法。利用特征加权融合算法跟踪动漫人物特征,将图像分成多个像素区域,根据像素损失值推导人物图像的变分损失复原函数。利用跟踪所得的人物特征,定义重建节点,再将各个节点对象连接起来,以求解具体的重建表达式,完成动漫人物图像复原重建。实验结果表明,应用构建模型可以保证复原图像中缺失像素与原像素的色彩一致性,色彩一致性高于92%,不会造成设计人物失真显示的情况,能够保障主机元件对动漫人物模型的准确重建。

The pixel loss in the restored image of anime characters can lead to distorted display of the designed characters,making it difficult for the host components to accurately reconstruct the anime character model.To solve the above problems,a feature weighted fusion algorithm based on animation character image restoration and reconstruction method is proposed.After the feature weighted fusion algorithm is used to track the animation characters,the image is divided into multiple pixel regions,and the variational loss restoration function is derived according to the pixel loss value.Using the tracked character features,define reconstruction nodes,and then connect each node object to solve the specific reconstruction expression to complete the restoration and reconstruction of anime character images.The experimental results show that the application of the constructed model can ensure the color consistency between the missing pixels and the original pixels in the restored images,and the color consistency is higher than 92%,which will not cause the distorted display of the design characters,and can ensure the accurate reconstruction of the animation character model by the host components.