Image Atlas File Encoding for Memory Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current image storage and transmission methods are inefficient, as they require significant memory resources and network bandwidth due to the need to store and transmit multiple images, especially when images have arbitrary shapes, leading to wasted storage capacity and high computational requirements.
Innovation Solution
A method and system for encoding target images into an image atlas file by identifying unique image elements, arranging them to minimize empty spaces, and generating metadata for efficient decoding and composition, reducing the size of the image atlas file and its metadata, thereby minimizing memory and network resource usage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If multiple images are stored and transmitted separately, then image quality and completeness are preserved, but memory resources and network bandwidth are consumed excessively
Solution Approach 1:
The patent merges multiple images into a single atlas by extracting common image elements and storing them once, then referencing them across multiple images through metadata. This reduces the total quantity of data stored while maintaining the ability to reconstruct complete images, directly addressing the contradiction between memory usage and image completeness.
Solution Approach 2:
The patent creates a universal image element library where common elements (like characters, objects, or textures) can be reused across multiple images. The same image element can serve multiple functions in different contexts, reducing redundancy and storage requirements while preserving image quality through selective reuse of universal elements.
2Quantity of substance
If images with arbitrary shapes are stored using standard rectangular formats, then storage simplicity is maintained, but storage capacity is wasted due to empty spaces
Solution Approach 1:
The patent segments images into discrete image elements that can be independently stored and referenced. By breaking down complex arbitrary-shaped images into smaller, manageable elements, the system can store only the necessary portions and arrange them efficiently in the atlas, reducing wasted storage space while maintaining the ability to reconstruct complete images.
Solution Approach 2:
The patent transitions from storing complete rectangular images to storing a compact atlas of image elements with associated metadata that describes their positions and dimensions. This dimensional shift from image-level storage to element-level storage with spatial information allows efficient use of storage capacity while maintaining image reconstruction capability.
3Quantity of substance
If a large number of images are transmitted over the network, then image variety and content richness are improved, but network transmission bandwidth is overloaded
Solution Approach 1:
The patent combines multiple images into a single atlas transmission, reducing the total data volume that needs to be sent over the network. Instead of transmitting many separate images, the system transmits one compressed atlas containing all unique image elements plus metadata, significantly reducing bandwidth requirements while preserving image variety through the metadata references.
Solution Approach 2:
The patent performs preliminary processing to extract and organize common image elements into the atlas before transmission. By pre-processing and organizing the image data into a standardized format with metadata, the system optimizes the transmission process and reduces the actual data that needs to be sent over the network, improving bandwidth efficiency.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Disclosed is a method and encoder for encoding target images(s) into an image atlas file. Image elements present in the target images are identified. A set of unique image elements is created comprising each unique image element identified in the one or more target images and only one instance of each non-unique image element identified in the one or more target images as a corresponding unique image element. The unique image elements of the set are arranged into the image atlas file in a manner that empty spaces in the image atlas file are reduced. A metadata file is generated corresponding to the image atlas file, the metadata file comprising meta information of each target image, wherein meta information of a given target image is indicative of locations and dimensions of unique image elements in the image atlas file that are required to compose the given target image.