Depth video intra-frame intelligent coding method
A deep video and intelligent coding technology, which is applied in the field of deep learning and video coding, can solve the problems of low coding efficiency and lack of robustness in different scenarios, and achieve the effect of improving quality
Active Publication Date: 2019-11-29
TIANJIN UNIV
View PDF5 Cites 2 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
[0006] The focus of existing deep video coding research is how to construct prediction models based on the characteristics of depth video. Applying color video intelligent coding method to depth video will inevitably lead to a decline in coding efficiency
Method used
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment 2
[0064] The following diagrams illustrate the experimental results
[0065] This method is integrated into HTM16.2, and three 3D video coding standard test sequences are used for test experiments, namely: Balloons, Kendo and Newspaper. The experiments are set to a full intra-frame coding configuration, and the quantization parameter pairs are set to {25 / 34, 30 / 39, 35 / 42, 40 / 45}. This method uses the original HTM16.2 platform as a benchmark algorithm to demonstrate the effectiveness of the proposed method.
[0066] figure 2 The experimental results show that, compared with HTM, this method achieves bit rate savings on the three test sequences, and the average BDBR reduction value reaches 6.5%.
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More PUM
Login to View More Abstract
The invention discloses a depth video intra-frame intelligent coding method. The method comprises the following steps: constructing a variable-resolution predictive coding mode consisting of down-sampling and up-sampling; down-sampling the current depth LCU to reduce the size of the current depth LCU, obtaining a low-resolution depth block and performing low-resolution encoding; performing up-sampling on the coded low-resolution depth coding unit by using a color feature-assisted convolutional neural network, and extracting depth features and color features by using a residual coding unit; carrying out dimension reduction processing on the extracted features, and then carrying out feature fusion on the two features after dimension reduction to obtain final fusion features; adding the fusion feature and a discrete cosine interpolation filtering result, so that data in a training process is always a residual error between a predicted value and a real value; and implanting the above stepsinto 3D-HEVC to serve as a new intra-frame prediction mode, and performing rate distortion cost comparison with other intra-frame prediction modes to select an optimal prediction mode.
Description
technical field [0001] The invention relates to the fields of video coding and deep learning, in particular to an intelligent coding method within a deep video frame. Background technique [0002] 3D video has attracted widespread attention because it can provide users with an immersive three-dimensional experience. 3D-HEVC (3D version of the new generation high-efficiency video coding standard) is a 3D extended coding method of HEVC. In addition to coding the color video sequence of each viewpoint, it also needs to code the depth video sequence corresponding to each viewpoint. Depth video contains the depth and disparity information of the scene, reflecting the distance, depth and distribution of objects in the scene, and its coding performance directly affects the stereoscopic perception quality of the scene. Depth videos consist of large smooth regions and sharp boundaries. There is a lot of spatial redundancy in large-area smooth regions, and boundaries play an importa...
Claims
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More Application Information
Patent Timeline
Login to View More Patent Type & Authority Applications(China)
IPC IPC(8): H04N19/597H04N19/593H04N19/59H04N19/11H04N19/132H04N19/186G06K9/62
CPCH04N19/597H04N19/593H04N19/59H04N19/11H04N19/132H04N19/186G06F18/253
Inventor 雷建军刘晓寰侯春萍张凯明张静何景逸
Owner TIANJIN UNIV
PatSnap Eureka turns technology decisions into work you can execute. Powered by our Innovation Knowledge Graph, it runs expert workflows across engineering, life sciences, materials and intellectual property. Get your review-ready output in minutes.
