Epitome-Based Video Encoding for Compression Efficiency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video compression techniques, such as H.264, are inefficient in coding all types of images or sequences, as they rely on block-based prediction methods that do not fully utilize global image information, leading to suboptimal compression and increased bit rates.
Innovation Solution
A method that constructs and utilizes an epitome, a condensed representation of an image sequence, for intra prediction, allowing for improved compression by using the epitome to predict image blocks without requiring block alignment and reducing the need for transmitting the epitome itself, instead reconstructing it at the decoder.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If block-based prediction methods are used for video compression, then the coding process is simplified and computationally efficient, but the compression efficiency is reduced and bit rate increases
Solution Approach 1:
The patent creates an epitome, which is a condensed copy or representation of the image sequence containing essential texture and contour information. This epitome serves as a reference for prediction, allowing the decoder to reconstruct images with higher efficiency. The epitome is transmitted instead of full-resolution reference frames, reducing bit rate while maintaining compression efficiency.
Solution Approach 2:
The patent extracts only the most relevant constituent elements (texture and contour components) from the original image sequence to create the epitome. By taking out and transmitting only these essential features rather than complete image data, the system achieves better compression efficiency with reduced bit rate.
2Measurement precision
If the epitome is transmitted to the decoder, then prediction accuracy is improved, but the bit rate increases
Solution Approach 1:
The epitome contains localized essential features (texture and contour components) rather than complete global image information. This local quality approach transmits only the necessary predictive information at reduced resolution, improving prediction accuracy for essential structures while keeping bit rate low.
Solution Approach 2:
The patent transitions from transmitting full-resolution image data to transmitting a condensed epitome representation. This dimensional reduction transforms the problem from transmitting complete pixel information to transmitting extracted feature information, achieving better prediction accuracy per bit transmitted.
3Productivity
If block-based prediction is used, then the coding process is efficient, but global image information is not fully utilized
Solution Approach 1:
The patent performs preliminary extraction of essential image features (texture and contour components) to create the epitome before the actual coding process. This preliminary action prepares global image information in a condensed form that can be efficiently utilized during prediction, combining global information utilization with coding efficiency.
Solution Approach 2:
The epitome serves as a condensed copy that captures global image information in an efficient format. By creating and transmitting this epitome copy, the system enables the decoder to utilize global information from multiple frames without the computational overhead of processing complete high-resolution images.
Data Source
Figure 1~3
Figure 4~7
AI summary
The invention relates to a method for encoding a sequence of images. According to the invention, such a method implements the following steps, for at least one current image of the sequence: - constructing an epitome representative of the current image, from said sequence, the current image and the images of said sequence belonging to the same plane; - generating a signal including said constructed epitome, - generating information associated with said constructed epitome, said information being added to said generated signal or to another signal.