Related Experiment Video
Updated: Mar 1, 2026
07:12
Whole-cell Super-Resolution Imaging via DNA-PAINT on a Spinning Disk Confocal with Optical Photon Reassignment
Published on: January 6, 2026
465
Segmented targeted least squares estimator for material decomposition in multibin photon-counting detectors
Paurakh L Rajbhandary1,2, Scott S Hsieh1, Norbert J Pelc1,2,3
1Stanford University, Department of Radiology, Palo Alto, California, United States.
Journal of Medical Imaging (Bellingham, Wash.)
|June 1, 2017
Summary
A new targeted least squares estimator (TLSE) offers fast and accurate material separation for photon-counting x-ray detectors (PCXDs). This method improves upon existing techniques, providing results comparable to maximum likelihood estimation but with significantly greater computational efficiency.
Area of Science:
- Medical Physics
- Biomedical Imaging
- Signal Processing
Background:
- Photon-counting x-ray detectors (PCXDs) offer improved spectral information compared to conventional detectors.
- Accurate material separation is crucial for quantitative imaging in various applications.
- Existing material separation methods may face limitations in speed, accuracy, or noise efficiency.
Purpose of the Study:
- To develop and evaluate a novel, fast, noise-efficient, and accurate estimator for material separation using PCXDs.
- To improve upon the A-table method by incorporating dynamic weighting for enhanced performance.
- To compare the proposed estimator against established methods like maximum likelihood estimation (MLE) and the A-table method.
Main Methods:
- The targeted least squares estimator (TLSE) was developed, incorporating dynamic weighting for variance optimization.
- Both Cartesian and average-energy segmentation strategies were explored for basis material space.
- Monte Carlo simulations were used to assess variance and bias across a range of material thicknesses (0-6 cm Al, 0-50 cm H2O).
Main Results:
- Average-energy TLSE achieved variance within 2% of the Cramér-Rao lower bound (CRLB) and a mean absolute error of [Formula: see text].
- TLSE demonstrated lower variance than the A-table method, particularly in peripheral operating ranges (thin or thick objects).
- TLSE was found to be approximately 50 times faster than MLE while achieving comparable accuracy and precision.
Conclusions:
- The targeted least squares estimator (TLSE) provides a computationally efficient and fast solution for material separation with PCXDs.
- TLSE offers accuracy and precision comparable to the gold standard MLE, with improved performance over the A-table method.
- The average-energy segmentation approach within TLSE requires fewer segments for similar performance compared to Cartesian segmentation.
