Object-Aware AI Image Compression for Storage Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The rising costs of storing visual media due to high-quality image and video storage demands, particularly in personal computing devices and cloud storage, are exacerbated by the need to maintain image quality and resolution for enjoyment and usefulness.
Innovation Solution
Implementing machine learning models to identify and classify objects in images, associate tags, and apply compression engines to create a compression pyramid, ensuring important features remain discernible while reducing storage space by selectively compressing backgrounds.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If full quality and resolution storage is maintained, then image enjoyment and usefulness are preserved, but storage costs rise dramatically
Solution Approach 1:
The patent applies different compression quality levels to different regions of the image. Important objects identified by machine learning models are preserved at high quality, while background regions are compressed more aggressively. This resolves the contradiction by maintaining image quality locally where it matters (for important objects) while reducing overall storage space through selective compression of less important areas.
2Quantity of substance
If compression is applied to reduce storage space, then storage costs decrease, but image quality and discernibility of important features may deteriorate
Solution Approach 1:
The patent performs preliminary object detection and classification using machine learning models before applying compression. By identifying important objects in advance, the system can pre-determine which regions require high-quality preservation and which can be compressed. This preliminary action ensures that compression is applied intelligently, maintaining quality where needed while achieving storage reduction.
Solution Approach 2:
The system uses machine learning models to analyze compressed images and verify that important objects remain discernible. This feedback mechanism allows the system to adjust compression parameters to ensure quality thresholds are met while maximizing storage reduction. The feedback loop ensures that compression does not degrade the discernibility of important features.
3Reliability
If machine learning models are used to identify and classify objects, then selective compression can be applied to maintain important features, but device complexity increases
Solution Approach 1:
The machine learning models automatically perform object detection, classification, and tagging without requiring manual intervention. The system self-services by autonomously identifying important regions and applying appropriate compression parameters. This automation reduces the need for complex manual configuration and processing, making the system more practical despite the underlying complexity of the AI models.
Data Source
AI summary
In non-limiting examples of the present disclosure, systems, methods and devices for determining image compression optimums are provided. An image may be processed with a machine learning model that has been trained to identify object types in digital images. A first object and a first object type of the first object may be identified in the image. A first compressed version of the image may be generated, wherein the first compressed version has a first storage size. The first object and the first object type of the first object may be identified in the first compressed version of the image. A second compressed version of the image may be generated based on the identification of the first object and the first object type in the first compressed version of the image. The second compressed version may have a smaller storage size than the first storage size.