Abstract
This chapter introduces the model-based iterative reconstruction (MBIR) framework and demonstrates how model-based approaches to electron tomography can dramatically improve reconstruction quality by reducing noise and missing wedge artifacts while being robust to outliers and missing calibration parameters. It outlines the MBIR paradigm and presents MBIR methods for different modalities in electron microscopy. The chapter provides some important future directions for MBIR approaches. MBIR is a systematic approach to solving inverse problems in imaging. The chapter discusses MBIR method for high angle annular dark field-scanning transmission electron microscope (HAADF-STEM) tomography. It examines a forward model for the measurement process in HAADF-STEM tomography and explores details of the prior model for the three-dimensional volume. The chapter focuses on MBIR algorithm for accurate reconstruction of bright field-electron tomography (BF-ET) data containing anomalous measurements that typically result from crystalline samples. The goal of BF-ET is to reconstruct an attenuation coefficient at every point in the sample.
| Original language | English |
|---|---|
| Title of host publication | Statistical Methods for Materials Science |
| Subtitle of host publication | The Data Science of Microstructure Characterization |
| Publisher | CRC Press |
| Pages | 85-109 |
| Number of pages | 25 |
| ISBN (Electronic) | 9781498738217 |
| ISBN (Print) | 9781315121062 |
| DOIs | |
| State | Published - Jan 1 2019 |