STITCHING LATERALLY ADJACENT TOMOGRAMS

这个子项目是许多研究子项目中利用 资源由NIH/NCRR资助的中心拨款提供。子项目和 调查员(PI)可能从NIH的另一个来源获得了主要资金， 并因此可以在其他清晰的条目中表示。列出的机构是 该中心不一定是调查人员的机构。 即使是通过剪辑重叠的相机帧拍摄的倾斜序列，在一张塑料切片的单张断层图像中重建的区域大小也有实际限制。为了重建更大的区域，有必要从相邻的重叠区域生成断层图像，并将它们缝合在一起。这项工作的主要推动力是可见细胞项目，该项目旨在重建胰腺细胞的很大一部分。在这个项目中，确定了可以在3x3重叠断层图像阵列中获得必要的区域，称为超级蒙太奇(每个断层图像本身由2Kx2K相机帧的3x3蒙太奇生成)。拼接这些断层图像的战略是在可能的情况下调整IMOD的现有工具，并在必要时开发新的工具。缝合是通过运行三个新程序来执行的。第一程序建立用于在处理序列中携带关于超级蒙太奇的信息的文件，找到对准每对相邻重叠体积所需的总体移位，并在每个重叠区域中的位置阵列处互相关小子体积，以测量与完美对准的局部偏差。第二个程序获取有关相邻卷对之间关系的信息，并将其解析为最佳对齐所有卷的转换。这些变换包括重叠区域中体积的一些非线性翘曲。这种变形的主要目标是将3D拼接问题减少到2D拼接问题，以便通过现有的程序混合2D蒙太奇图像来完成最终的拼接。第三个程序管理该过程的最后步骤：扭曲每个单独的体积，匹配体积之间的图像强度，将它们堆叠到一个文件中，并使用一些适合这种情况的特殊参数运行剪辑混合程序。有了这些程序，一个3x3的超级蒙太奇已经被无缝缝合在一起。

This subproject is one of many research subprojects utilizing the resources provided by a Center grant funded by NIH/NCRR. The subproject and investigator (PI) may have received primary funding from another NIH source, and thus could be represented in other CRISP entries. The institution listed is for the Center, which is not necessarily the institution for the investigator. There are practical limits to the size of the area that can be reconstructed in a single tomogram from a plastic section, even with tilt series taken by montaging overlapping camera frames. To reconstruct larger areas, it is necessary to generate tomograms from adjacent, overlapping areas and stitch them together. The primary impetus for this effort is the Visible Cell Project, which aims to reconstruct a significant fraction of a pancreatic cell. For this project, it was determined that the necessary area could be acquired in a 3x3 array of overlapping tomograms, referred to as a supermontage (each tomogram is itself generated from a 3x3 montage of 2Kx2K camera frames). The strategy for stitching these tomograms was to adapt existing tools in IMOD where possible and develop new ones where necessary. Stitching is carried out by running three new programs. The first program sets up a file for carrying information about the supermontage through the processing sequence, finds the overall shifts needed to align each pair of adjacent overlapping volumes, and cross-correlates small subvolumes at an array of positions in each overlap zone to measure the local deviations from perfect alignment. The second program takes this information about the relationships between adjacent pairs of volumes and resolves it into transformations that will best align all of the volumes. These transformations include some nonlinear warping of the volumes in the overlap zones. The primary goal of this warping is to reduce the 3-D stitching problem to a 2-D one so that the final stitching can be done by the existing program from blending 2-D montaged images. The third program manages the final steps of the process: warping each individual volume, matching image intensities between volumes, stacking them into a single file, and running the montage-blending program with some special parameters appropriate for this situation. With these programs, a 3x3 supermontage has been stitched seamlessly.