Fully automatic and robust 3D registration of serial-section microscopic images.

Ching-Wei Wang,Eric Budiman Gosno,Yen-Sheng Li

doi:10.1038/srep15051

Abstract

Robust and fully automatic 3D registration of serial-section microscopic images is critical for detailed anatomical reconstruction of large biological specimens, such as reconstructions of dense neuronal tissues or 3D histology reconstruction to gain new structural insights. However, robust and fully automatic 3D image registration for biological data is difficult due to complex deformations, unbalanced staining and variations on data appearance. This study presents a fully automatic and robust 3D registration technique for microscopic image reconstruction, and we demonstrate our method on two ssTEM datasets of drosophila brain neural tissues, serial confocal laser scanning microscopic images of a drosophila brain, serial histopathological images of renal cortical tissues and a synthetic test case. The results show that the presented fully automatic method is promising to reassemble continuous volumes and minimize artificial deformations for all data and outperforms four state-of-the-art 3D registration techniques to consistently produce solid 3D reconstructed anatomies with less discontinuities and deformations.

Highlights

Robust and fully automatic three-dimensional (3D) image registration of serial-section microscopic images is critical for detailed 3D anatomical reconstruction of large biological specimens such as serial section Transmission Electron Microscopy of neural tissues[1], serial confocal laser scanning microscopic images of a brain[2,3] or serial histopathological microscopic images[4,5]
The main contribution of this study is to present a fully automatic and robust 3D image registration method for reconstruction of detailed 3D anatomy and able to deal with complex deformation problems for different types of microscopic images, such as serial serial section Transmission Electron Microscopy (ssTEM) images, laser scanning confocal images and histopathological images
Four state-of-the-art 3D registration methods are compared with the proposed method using TrakEM21,7–9, including a method using least squares[16], an elastic b-spline model for biological images (UnwarpJ)[17], an improved bi-directional b-spline model for histopathological section alignment[18] and an elastic volume reconstruction method[1], and the four benchmark registration methods are tested with four different transformation parameters, including translation, rigid, similarity and affine

Summary

Introduction

Robust and fully automatic three-dimensional (3D) image registration of serial-section microscopic images is critical for detailed 3D anatomical reconstruction of large biological specimens such as serial section Transmission Electron Microscopy (ssTEM) of neural tissues[1], serial confocal laser scanning microscopic images of a brain[2,3] or serial histopathological microscopic images[4,5]. In comparison to laser scanning confocal images as used in the studies[2,3] where the serial image data maintains the property of geometrical continuity in 3D space, there are complex deformation problems for serial histopathological slides, including physical destructions caused by cutting and fixation, staining artifacts and uneven stain variations due to potential discrepancy in thickness of individual tissue sections. These complex distortion effects makes image registration of histopathological data an even harder task. The experimental results show that the proposed fully automatic method is promising to perform 3D registration well for all data and consistently produces solid 3D reconstructed objects with less discontinuities and deformations in comparison to the benchmark methods

Objectives

Methods

Results

Conclusion