Technical Description
The technology discloses a fully unsupervised image-reconstruction framework combining deep learning with iterative reconstruction. It calculates projection-data error, updates neural-network parameters through a loss function, and maps reconstructed images back into projection space.
Problems Addressed
- Reconstruction Speed-Quality Trade-Offs
- Reconstruction Efficiency Limitations
- Noise & Blur Artifacts
- Truncated Projection Data
- Prior Truncation Correction Requirements
- Regularization-Based Solution Restrictions
Tech Features
- Deep Learning & Iterative Reconstruction
- Projection-Error Feedback
- Loss-Driven Parameter Updates
- Radon Transform Mapping
- Iterative Optimum-Loss Search
- Truncated-Data Reconstruction
Target Audience
- Biotechnology companies
- Healthcare and medical device industry
- Analytical chemistry laboratories
- Point-of-care diagnostics developers
- Research and development laboratories
Tech ID: P02-2310 TRL 4 Patent Status: Granted Available For Exclusive and Non-exclusive License
