Medical Image Compression via Seed Segments
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current lossless image compression algorithms for medical images do not provide a satisfactory compression ratio, leading to high storage costs due to the large size of DICOM images generated by medical imaging modalities, which are often stored for extended periods, necessitating a more effective method to reduce storage requirements without losing pixel data.
Innovation Solution
A method involving image segmentation, comparison with seed images, and replacement of non-matching segments with seed image identifiers, allowing for the generation and storage of residual and seed images, which are used for efficient compression and reconstruction of medical images.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If lossless compression algorithms are applied to medical images, then pixel data is preserved without loss, but compression ratio is insufficient leading to high storage costs
Solution Approach 1:
The medical image is divided into multiple segments or blocks. Each segment is independently processed to identify and replace redundant regions with references to seed images, enabling localized compression while preserving overall image fidelity.
Solution Approach 2:
Seed images are created from representative segments and stored separately. Instead of storing duplicate segment data multiple times, the patent uses copy references (seed image identifiers) to point to these master copies, significantly reducing storage requirements while maintaining data integrity.
2Reliability
If current lossless compression algorithms are used, then data integrity is maintained, but storage costs escalate quickly over time
Solution Approach 1:
Multiple similar or identical image segments are merged into a single seed image. The storage system stores one master copy (the seed image) and references to it, rather than storing each segment separately, thereby consolidating storage requirements and reducing costs.
Solution Approach 2:
The patent changes the storage parameter from storing complete image data for every segment to storing compact references (seed image identifiers). This parameter change from full data storage to reference storage dramatically reduces storage costs while maintaining data integrity through lossless reconstruction.
3Productivity
If image segmentation and seed image association are implemented, then compression ratio is improved, but algorithm complexity increases
Solution Approach 1:
Seed images are pre-processed and stored in a seed image database before main compression operations. This preliminary organization of reference images simplifies the subsequent compression process by providing ready-to-use templates for comparison and reference, reducing the computational complexity of the main algorithm.
Data Source
AI summary
The disclosure relates to method and system for compressing an image. The method involves receiving an image of the one or more images. Further, at least one segmentation algorithm is applied on the image and dividing the image into a plurality segments. The method further includes comparing the plurality of segments of the image with a seed image, where seed images include a seed image identifier. Further, a seed image is associated with the segments of the image in case there is a match between the seed image and the plurality of segments. The method also includes storing the image as a residual image and a seed image along with one or more seed image identifiers. Further, the image may be reconstructed based on the residual image and one or more seed images associated with the image. Thereafter, the image may be displayed on a display unit.