Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Automatic Registration of SHG and H&E Images with Feature-based Initial Alignment and Intensity-based Instance Optimization: Contribution to the COMULIS Challenge

Marek Wodzinski, Henning Müller

TL;DR

This work proposes a method based on automatic keypoint matching followed by deformable registration based on instance optimization that achieved relatively good generalizability resulting in 88% of success rate in the initial alignment and average target registration error equal to 2.48 on the external validation set.

Abstract

The automatic registration of noninvasive second-harmonic generation microscopy to hematoxylin and eosin slides is a highly desired, yet still unsolved problem. The task is challenging because the second-harmonic images contain only partial information, in contrast to the stained H&E slides that provide more information about the tissue morphology. Moreover, both imaging methods have different intensity distributions. Therefore, the task can be formulated as a multi-modal registration problem with missing data. In this work, we propose a method based on automatic keypoint matching followed by deformable registration based on instance optimization. The method does not require any training and is evaluated using the dataset provided in the Learn2Reg challenge by the COMULIS organization. The method achieved relatively good generalizability resulting in 88% of success rate in the initial alignment and average target registration error equal to 2.48 on the external validation set. We openly release the source code and incorporate it in the DeeperHistReg image registration framework.

Automatic Registration of SHG and H&E Images with Feature-based Initial Alignment and Intensity-based Instance Optimization: Contribution to the COMULIS Challenge

TL;DR

This work proposes a method based on automatic keypoint matching followed by deformable registration based on instance optimization that achieved relatively good generalizability resulting in 88% of success rate in the initial alignment and average target registration error equal to 2.48 on the external validation set.

Abstract

The automatic registration of noninvasive second-harmonic generation microscopy to hematoxylin and eosin slides is a highly desired, yet still unsolved problem. The task is challenging because the second-harmonic images contain only partial information, in contrast to the stained H&E slides that provide more information about the tissue morphology. Moreover, both imaging methods have different intensity distributions. Therefore, the task can be formulated as a multi-modal registration problem with missing data. In this work, we propose a method based on automatic keypoint matching followed by deformable registration based on instance optimization. The method does not require any training and is evaluated using the dataset provided in the Learn2Reg challenge by the COMULIS organization. The method achieved relatively good generalizability resulting in 88% of success rate in the initial alignment and average target registration error equal to 2.48 on the external validation set. We openly release the source code and incorporate it in the DeeperHistReg image registration framework.
Paper Structure (11 sections, 1 equation, 6 figures, 1 table)

This paper contains 11 sections, 1 equation, 6 figures, 1 table.

Table of Contents

  1. Introduction
  2. Method
  3. Overview
  4. Preprocessing
  5. Initial Alignment
  6. Deformable Registration
  7. Dataset & Experimental Setup
  8. Results
  9. Discussion
  10. Acknowledgements
  11. Disclosure of Interests

Figures (6)

  • Figure 1: Exemplary pair of H&E and SHG images. Note the significantly different intensity distributions and the amount of missing data in the SHG image.
  • Figure 2: Visualization of the registration pipeline presenting also the intermediate results. Best viewed zoomed in. The target SHG image is presented in the original intensity range to emphasize the task difficulty (zoom required to see the details).
  • Figure 3: Visualization of the proposed preprocessing.
  • Figure 4: The TRE calculated for the validation pairs using the Grand-Challenge platform.
  • Figure 5: Qualitative registration results using several samples from the validation subset. The source and target images are overlayed in different color channels to present the alignment quality. The preprocessed images are used for the presentation clarity. Note the small impact of the deformable registration.
  • ...and 1 more figures