Image Encoding Apparatus Object Map Prediction for Distance Images
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image encoding methods fail to efficiently encode distance images due to sharp edges in prediction residuals, which obstruct energy concentration in the frequency domain, leading to high-frequency components and poor encoding efficiency.
Innovation Solution
The method divides image frames into blocks, assigns object identification information and pixel values to each pixel, and generates an object map to predict pixel values, reducing the amount of code required for encoding by representing complex edges accurately.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional image encoding methods are used on distance images, then the encoding process is simple, but sharp edges in prediction residuals obstruct energy concentration in the frequency domain, leading to high-frequency components and poor encoding efficiency
Solution Approach 1:
The image is divided into multiple blocks, and each block is further segmented into multiple regions based on depth value characteristics. This segmentation allows different encoding strategies to be applied to different regions, enabling accurate representation of edges while maintaining encoding efficiency.
Solution Approach 2:
Different regions within blocks are encoded with different precision levels based on their local characteristics. Regions with significant depth variations (edges) are encoded with higher precision, while uniform regions use lower precision, optimizing the balance between quality and compression ratio.
2Manufacturing precision
If the image is divided into blocks and object identification information is assigned to each pixel, then complex edges can be represented accurately, but the device complexity increases
Solution Approach 1:
The encoding apparatus segments the image into blocks and further divides blocks into regions, applying different encoding processes to each segment. This structured segmentation manages complexity by breaking down the overall complex task into simpler, manageable sub-tasks.
Solution Approach 2:
The encoding process dynamically adapts to local image characteristics by automatically determining region boundaries and selecting appropriate encoding parameters for each region, reducing the need for manual configuration and simplifying the overall system complexity.
3Loss of information
If object maps are generated to predict pixel values, then the code required for encoding is reduced, but the processing time increases
Solution Approach 1:
The image processing is segmented into blocks and regions, allowing object map generation to be performed locally and efficiently. This segmentation enables parallel processing of different blocks, reducing overall processing time while maintaining compression efficiency.
Solution Approach 2:
Object maps are generated in advance during the encoding process and stored for reuse in prediction operations. This preliminary generation of prediction data eliminates the need for repeated complex calculations during decoding, reducing processing time.
Data Source
AI summary
An image encoding method in which when transmitting or storing an image, a frame of the image is divided into predetermined-sized processing regions, and for each processing region, a pixel value of each pixel is predicted for the encoding. The method includes a step that determines one pixel value, which is assigned to and represents each object in the processing region, to be an object pixel value that is associated with an object identifier for identifying the relevant object; a step that generates, based on each object pixel value and the pixel value of each pixel in the processing region, an object map that indicates which object has been obtained at each pixel in the processing region, by using the object identifier; a step that generates a predicted image for the processing region by assigning the object pixel value to each pixel in accordance with the object map; a step that encodes the object map; a step that encodes each object pixel value; and a step that performs predictive encoding of an image signal for the processing region by using the predicted image.