3D Compressed Sensing With Dictionary Learning for Depth Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems face challenges in efficiently transmitting and utilizing three-dimensional data in real time due to increased data volume and bandwidth constraints, leading to inaccuracies in distance and depth maps, making it difficult to construct IoT systems that handle three-dimensional data without further processing.
Innovation Solution
A compressed sensing system that uses dictionary learning to partially reduce data on the transmission side and restore it on the reception side based on basis matrices and sparsity, employing additional basis vectors for accurate reconstruction of depth and color information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If three-dimensional data is transmitted in full resolution, then measurement precision is maintained, but data communication volume increases and bandwidth is tightened
Solution Approach 1:
The patent extracts only the essential features needed for three-dimensional reconstruction by performing principal component analysis to identify and retain only the most significant components, discarding redundant information while preserving measurement precision
Solution Approach 2:
The patent transforms the data representation by changing from transmitting full-resolution three-dimensional data to transmitting compressed spectral data in the frequency domain, using parameter transformation through Fourier transform and selective component retention
2Quantity of substance
If two-dimensional compression methods are applied to three-dimensional data, then data communication volume is reduced, but manufacturing precision deteriorates due to frequency concentration loss
Solution Approach 1:
The patent transitions from two-dimensional image compression to three-dimensional spectral domain compression by introducing the frequency dimension through Fourier transform, allowing compression in the spectral domain while preserving spatial reconstruction accuracy
Solution Approach 2:
The patent creates a universal compression framework that handles both depth information and color information through the same spectral compression mechanism, enabling multi-functional processing of different three-dimensional data types with a single method
3Quantity of substance
If data is decimated and transferred using conventional filtering, then data communication volume is reduced, but measurement precision is lost due to low-frequency concentration
Solution Approach 1:
The patent applies different retention strategies to different frequency components, preserving high-frequency components that contain critical distance and depth information while allowing compression of less critical low-frequency components, ensuring local quality preservation where needed
Data Source
AI summary
Three-dimensional sensing data is transferred efficiently by applying compressed sensing using dictionary learning. Conventional image-based lossy compression improves a feature according to which power is concentrated on low-frequency components and deterioration of three-dimensional information increases. By changing the decimation rates for the depth information and the color information and using a dictionary vector created from the other restoration result for restoration, the original result is reconstructed using a small data volume.