基于深度学习的分形图像压缩编码模型设计

Design of fractal image compression coding model based on deep learning

针对传统编码模型存在的图像分割处理效果不佳、分割精度不高、耗时较长以及编码质量较差的问题,提出基于深度学习的分形图像压缩编码模型。建立图像分割约束条件,对图像进行分割处理,以处理后的图像作为深度学习中ResNet网络模型的输入,提取原始分形图像的特征并将图像块分类;建立特征图像块匹配规则,排序图像块,记录分形码,将分形码作为图像在度量空间内的表现形式,通过度量空间的压缩变换实现分形图像压缩编码。实验结果表明:所提模型的图像压缩编码质量较高,具有一定的应用价值。

In view of the poor image segmentation processing effect of traditional coding models, low segmentation accuracy, long time-consuming and poor coding quality, a fractal image compression coding model based on deep learning is proposed. The image segmentation constraint is established to segment the image, and the processed image is used as the input of the ResNet network model in deep learning to extract the features of the original fractal image and classify the image block. Then, the feature image block matching rules are established, the image blocks are sorted and the fractal codes are recorded which would be used as the manifestation of the image in the metric space, and the fractal image compression coding is finally realized through the compression transformation of the metric space. The experiment results show that the image compression coding quality of the proposed model is high, it has certain application value.