This invention is patented. Patent number is VN-0021155
This patent mentions to a novel distributed video encoding and decoding method, namely DVC-JEM, where video frames are separately encoding but jointly decoding using the recent Joint Exploration Model (JEM) codec. First, the video sequence is separated into the KEY and WYNER-ZIV (WZ) frames (100). Then, the KEY frames are encoded and decoded using the JEM video codec (200 and 300) while the WZ frames are encoded and decoded using the WZ coding approach with: a discrete cosine transform (400), a quantization module (500), Low density parity Codes module (600), side information (SI) creation (900), correlation noise model (1000), reconstruction (1100) and inverse discrete cosine transform (1200). The invention provides a process of encoding and decoding a video as a distributed manner to achieve low computation complexity at the encoder side. Figure 1 detailed the DVC-JEM encoder and decoder components.
The patent DVC-JEM provided a superior compression performance and lower complexity requirement when compared to relevant video coding methods, including the standard H.264/AVC and H.265/HEVC (see Figure 2). The mentioned patent is suitable for wireless video sensor networks and smart surveillance visual system applications in which the low complexity requirement is an important aspect.
Novelty:
– The invention mentions a novel method for distributed encoding and decoding videos using the JEM solution;
– The inventions mentions a process of coding video with side information and correlation noise estimation at the decoder side;
Creativity:
– The use of newly developed JEM solution for KEY frame coding;
– The creation of an estimated original WZ video information at the decoder side using the decoded KEY frames
Useful & Application capability:
– The DVC-JEM video coding method achieved lower computation complexity than standard H.264/AVC and HEVC solution; thus, it is suitable for a large range of video applications where low complexity requirement at the encoder side is vital such as: wireless sensor networks, surveillance visual systems, remote sensing.
