Tensor Display Light Field Compression Using Deformable SAI Embedding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing light field compression methods for multi-layer 3D displays, such as tensor displays, are inefficient and result in poor pixel estimates due to their highly redundant nature and sub-optimal compression schemes, particularly when using least-squares algorithms.
Innovation Solution
A machine learning model, specifically a deep learning deformable SAI feature embedding network, is trained to optimize pixel representations for each layer of a 3D display, utilizing deformable convolution and residual learning to minimize loss functions, enabling efficient compression and encoding aligned with the properties of multi-layer displays.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If least-squares algorithm is used to compress light field data for tensor display, then compression is achieved, but pixel estimates are poor and quality deteriorates
Solution Approach 1:
The patent transforms the compression approach by changing from direct pixel value compression using least-squares to compressing angular frequency domain representations. This parameter transformation in the frequency domain preserves essential light field information while achieving compression, then reconstructs high-quality pixel estimates through inverse transformation.
Solution Approach 2:
The patent introduces angular frequency domain representations as an intermediary between the original light field data and the compressed form. This intermediary representation captures the essential structure of light field data more efficiently, allowing for better compression while maintaining or improving pixel estimate quality.
2Productivity
If sequence-based compression of Synthetic Aperture Image is employed, then some compression is achieved, but it is inefficient for tensor display-based presentation requiring 169 pieces of information for 17×17 SAI
Solution Approach 1:
The patent moves from spatial domain compression to angular frequency domain compression, effectively changing the dimension of analysis. This transformation allows the compression algorithm to exploit correlations in the angular frequency domain that are not apparent in the spatial domain, achieving better compression ratios with reduced complexity for tensor display.
3Productivity
If machine learning model is trained to optimize pixel representations for each layer, then computing and network resources are reduced, but model training and inference complexity increases
Solution Approach 1:
The patent applies machine learning models in advance during the compression stage to learn optimal representations of light field data. The model is trained beforehand to capture the statistical regularities and correlations in light field data, enabling efficient compression and reconstruction without requiring complex real-time processing during deployment.
Data Source
AI summary
Systems, methods and apparatuses are described herein for training a machine learning model to accept as input synthetic aperture image (SAI) training data for a three-dimensional (3D) display, the 3D display comprising a plurality of layers. The machine learning model may be trained to output respective pixel representations of the SAI training data for each of the plurality of layers of the 3D display. The provided systems, methods and apparatuses may access image data, input the image data to the trained machine learning model, and determine, using the trained machine learning model, respective pixel representations of the input image data for each of the plurality of layers of the 3D display. The provided systems, methods and apparatuses may encode the respective pixel representations of the input image data, and transmit, for display at the 3D display, the encoded respective pixel representations of the input image data.