Face image super-resolution reconstruction method based on DCT domain eigen transform

Abstract

The invention relates to a face image super-resolution reconstruction method based on DCT (Discrete Cosine Transform) domain eigen transform, and relates to the image processing field. For the face image super-resolution reconstruction method, the energy concentration characteristic for DCT is utilized; the low frequency part utilizes a DCT domain interpolation amplification method; the high frequency part utilizes a method based on PCA (Principal Component Analysis) to predict the high frequency information; and as the input LR (Low Resolution) image is operated in the DCT domain, after processing, and a method based on bilateral filtering for the DCT domain is used, the blocking effect is removed adaptively. AA blocking DCT coefficient of the image can be obtained by decompressing the compressed image part; the blocking DCT coefficient is directly applied to a provided algorithm, so that the operation time for the image processing algorithm is reduced. For the compressed image, as rough quantification of the DCT coefficient causes distortion of image compression, super-resolution restoration is performed in the DCT domain so that algorithm process is operated from the source of distortion and the distortion problem of image compression can be preferably solved. Compared with a traditional algorithm, the face image super-resolution reconstruction method based on DCT domain eigen transform can be directly applied to a compressed image, and the reconstructed image can have higher subjective and objective quality.

Description

technical field [0001] The invention relates to an image processing method, in particular to a super-resolution reconstruction method of a human face image based on DCT domain eigentransformation. Background technique [0002] High-quality images and videos are increasingly becoming a mainstream demand because of their richer information and more realistic visual experience. Limited by the imaging environment and the performance of the imaging system, the images obtained by the imaging system are usually low-quality images with low definition. For example, many security departments, sensitive public places, traffic arteries, and residential areas are equipped with round-the-clock real-time video surveillance systems. However, due to factors such as the resolution performance of the surveillance camera, monitoring environment lighting conditions, monitoring distance, and noise, the video images acquired by the surveillance system may be blurred and low-quality images, and th...

AI Technical Summary

Problems solved by technology

However, when the magnification factor increases, the reconstruction constraints can provide less and le

Images