Interactive quantized noise calculating method in compressed video super-resolution

Abstract

The invention discloses an interactive quantized noise calculating method in compressed video super-resolution. The method includes the following steps of: first counting the appearance probability of DCT coefficients before quantization before a coding end does quantized operation to the DCT coefficients of a video frame and calculating Laplace parameter capable of representing the distribution of the DCT coefficients before quantization; then writing the distribution parameter of the DCT coefficients before quantization of an obtained image block in a Data-user field reserved in a code stream to be sent to a decoding end through coding; and finally obtaining the distribution parameter of the DCT coefficients before quantization in the code stream from the decoding end, calculating and obtaining the quantized noise according to the distribution probability density of the DCT coefficients before quantization and coefficients after quantization and consequently obtaining a final high-resolution image in a super-resolution algorithm. The calculating method is applicable to the compressed video super-resolution algorithm interacting between the coding end and the decoding end, improves the accuracy of quantized noise by obtaining the distribution parameter of the DCT coefficients before quantization at the coding end and providing the distribution parameter for the decoding end to calculate the quantized noise.

Description

technical field [0001] The invention relates to an interactive quantization noise calculation method in compressed video super-resolution, belonging to the field of multimedia communication. Background technique [0002] In video surveillance, for the needs of transmission and storage, the resolution of the image is often small. The mainstream of the general market is CIF (352*288) resolution, and the resolution of the video playback is often difficult to identify the monitoring object. The commonly used method to increase the resolution is to use better monitoring equipment, but this requires a large investment and the original equipment cannot be reused. Therefore, it is very urgent and necessary to improve the resolution of the surveillance video through software without adding new monitoring equipment. important. In the traditional super-resolution reconstruction algorithm, the reasons for image degradation, such as optical blur, motion blur, downsampling and noise, are...

AI Technical Summary

Problems solved by technology

In compressed video, quantization is the main cause of information loss during the compression process. In order to obtain an accurate quantization noise model, it is necessary to obtain the distribution characteristics of the DCT coefficients before quantization. The existing algorithms assume that the parameters of the DCT coefficient distribution model are either fixed parameters or Using it as a variable for iterative solution, the accuracy is not high, and there are also problems such as high computational complexity

The quantization noise modeling in the reconstruction method of traditional models for compressed video generally adopts the uniform distribution of DCT domain coefficients before quantization. Studies have shown that the distribution of DCT coefficients before quantization satisfies the Laplace distribution, so the traditional super-resolution method for compressed video The DCT coefficient distribution model before quantization in the high-rate reconstruction algorithm cannot accurately express the quantization noise, so it will affect the effect and quality of ultra-compressed video super-resolution reconstruction

On the other hand, simply reconstructing super-resolution images on the decoder side has two problems: (1) The distribution of DCT coefficients before quantization conforms to the Laplace distribution, but the decoder side only has quantized DCT coefficients, and the calculation of quantization noise requires the introduction of new variables to represent the parameters of the Laplace distribution, and the value of this parameter is obtained through iterative optimization, which increases the computational overhead; (2) Quantization noise is characterized and calculated in the DCT domain, and frequency domain constraints need to be introduced in the super-resolution reconstruction algorithm, High computational complexity and slow convergence

Images