Abstract:BACKGROUND: A coded aperture X-ray diffraction (XRD) imaging system can measure the X-ray diffraction form factor from an object in three dimensions –X, Y and Z (depth), broadening the potential application of this technology. However, to optimize XRD systems for specific applications, it is critical to understand how to predict and quantify system performance for each use case. OBJECTIVE: The purpose of this work is to present and validate 3D spatial resolution models for XRD imaging systems with a detector-side coded aperture. METHODS: A fan beam coded aperture XRD system was used to scan 3D printed resolution phantoms placed at various locations throughout the system’s field of view. The multiplexed scatter data were reconstructed using a model-based iterative reconstruction algorithm, and the resulting volumetric images were evaluated using multiple resolution criteria to compare against the known phantom resolution. We considered the full width at half max and Sparrow criterion as measures of the resolution and compared our results against analytical resolution models from the literature as well as a new theory for predicting the system resolution based on geometric arguments. RESULTS: We show that our experimental measurements are bounded by the multitude of theoretical resolution predictions, which accurately predict the observed trends and order of magnitude of the spatial and form factor resolutions. However, we find that the expected and observed resolution can vary by approximately a factor of two depending on the choice of metric and model considered. We observe depth resolutions of 7–16 mm and transverse resolutions of 0.6–2 mm for objects throughout the field of view. Furthermore, we observe tradeoffs between the spatial resolution and XRD form factor resolution as a function of sample location. CONCLUSION: The theories evaluated in this study provide a useful framework for estimating the 3D spatial resolution of a detector side coded aperture XRD imaging system. The assumptions and simplifications required by these theories can impact the overall accuracy of describing a particular system, but they also can add to the generalizability of their predictions. Furthermore, understanding the implications of the assumptions behind each theory can help predict performance, as shown by our data’s placement between the conservative and idealized theories, and better guide future systems for optimized designs.

Deriving a Quantitative Relationship Between Resolution and Human Classification Error

Distance v.s. Resolution: Neuromapping of Effective Resolution onto Physical Distance

Automatic Classifiers as Scientific Instruments: One Step Further Away from Ground-Truth

Influence of Thermal Imagery Resolution on Accuracy of Deep Learning based Face Recognition

Susceptibility to Image Resolution in Face Recognition and Trainings Strategies

Do humans and machines have the same eyes? Human-machine perceptual differences on image classification

Quantifying and mitigating the impact of label errors on model disparity metrics

Modeling the Trade-off of Privacy Preservation and Activity Recognition on Low-Resolution Images

Achieving Super-Resolution with Redundant Sensing

Methods of interpreting error estimates for grayscale image reconstructions

Beyond MR Image Harmonization: Resolution Matters Too

Using Relative Projection Density for Classification of Terrestrial Laser Scanning Data with Unknown Angular Resolution

Analysis of Machine Learning Prediction Reliability based on Sampling Distance Evaluation with Feature Decorrelation

Class-wise Autoencoders Measure Classification Difficulty And Detect Label Mistakes

Attention to detail: inter-resolution knowledge distillation

Effects of Digital Elevation Model Resolution on Unmanned Aerial Vehicle-Based Topographic Change Detection in Human-Altered Landscapes

Dynamic Amelioration of Resolution Mismatches for Local Feature Based Identity Inference

Predicting Performance of a Face Recognition System Based on Image Quality

A Hybrid Data Deduplication Approach in Entity Resolution Using Chromatic Correlation Clustering.

Neighborhood evaluator for efficient super-resolution reconstruction of 2D medical images

Resolution analysis of a volumetric coded aperture X-ray diffraction imaging system