Adaptive Lossy Compression Using Machine Learning Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current lossy compression techniques do not effectively adapt to the informational value and context-specific importance of media data, leading to inefficient storage and access of image, audio, and video data.
Innovation Solution
The use of machine learning models to infer criteria for lossy compression based on user inputs and contextual features, allowing for adaptive decision-making on whether to downsample, compress, keep, or remove specific features or frames of media data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If conventional lossy compression techniques are used to reduce storage size and bandwidth, then the storage efficiency is improved, but the informational value and context-specific importance of media data are lost
Solution Approach 1:
The patent applies different compression strategies to different regions or features of media data based on their informational value. Machine learning models identify and prioritize important features (such as key objects, faces, or semantically significant regions) while applying more aggressive compression to less important areas, thereby maintaining storage efficiency while preserving critical information.
Solution Approach 2:
The system dynamically adjusts compression parameters based on the informational value of different media features. Machine learning models analyze media data to determine which features warrant higher preservation quality and modify compression settings accordingly, changing parameters such as compression ratio, bitrate, or resolution selectively across different portions of the media.
2Productivity
If conventional lossy compression techniques are used to reduce bandwidth, then the transmission efficiency is improved, but the contextual importance of media features is compromised
Solution Approach 1:
The patent transmits media data with differentiated quality levels based on contextual importance. Machine learning models identify high-value features and allocate more bandwidth to their transmission while reducing bandwidth for less important content, optimizing transmission efficiency without losing critical contextual information.
Solution Approach 2:
The system uses machine learning models to continuously analyze media content and provide feedback on which features are contextually important. This feedback loop enables dynamic adjustment of transmission parameters in real-time, ensuring that bandwidth is allocated to preserve the most valuable information while maintaining overall transmission efficiency.
3Adaptability or versatility
If machine learning models are used to infer compression criteria based on informational value, then the adaptability to context-specific importance is improved, but the system complexity increases
Solution Approach 1:
The patent introduces machine learning models as intermediary components that bridge the gap between raw media data and compression algorithms. These models act as intelligent mediators that analyze media content, determine informational value, and generate compression criteria, thereby enabling context-specific adaptability while managing system complexity through modular architecture.
Solution Approach 2:
The system performs preliminary analysis of media data using machine learning models to infer compression criteria before the actual compression process. By pre-processing media content to identify important features and determine appropriate compression parameters in advance, the system achieves high adaptability while organizing complexity into separate, manageable stages.
4Loss of information
If machine learning models dynamically adjust compression criteria, then the informational value preservation is improved, but the processing time increases
Solution Approach 1:
The patent performs machine learning-based analysis and compression criteria determination as a preliminary step before actual compression. By inferring compression parameters in advance based on informational value assessment, the system preserves critical information while organizing processing into efficient stages that minimize overall processing time.
Solution Approach 2:
The system applies machine learning analysis selectively to portions of media data that require high informational value preservation, rather than uniformly processing all content. By focusing computational resources on critical features and applying simpler compression to less important areas, the system maintains information quality while reducing overall processing time.
Data Source
AI summary
Systems and methods are provided for obtaining a media, the media including an image, audio, video, or combination thereof. An input may be received regarding one or more features or frames of the media to be maintained in or removed from the media. One or more criteria of a lossy compression technique may be inferred based on the received input, using a machine learning model, based on the received input. The inferred criteria of the lossy compression technique may be applied to the media.