Here are a few sections at 1 micron.

Why these sections? Because these are some of the sections used by Konrad Wagstyl as a training set for the automated segmentation of the cortical layers using deep learning.

Other sections can be made available upon demand, but we cannot provide all of them due to lack of resources.

Some sections are in BigTiff format, some have been converted to minc . All tif images use 8-bit encoding; some are stored as 8 bits while the others are stored as 16 bits.

A converter bigtiff2mnc.c is supplied for the conversion of BitTiff to minc. The instructions to compile it are in the header of the source file. You need to install the BigTiff library and some basic minc tools. Use at your own risks.

For viewing the sections in bigtiff format, one can install the LargeTIFFTools. Alternately, you can download the HuronViewer.

Also, note that the 1 micron sections have been rescanned at a much later time than the scan time of the original sections at 10 microns (flatbed scanner). This means that the 1 micron sections are not aligned to the previous acquisitions. Furthermore, the quality of the images is also degraded due to stain crystallization. The specifiers TS01, TS02, TS03 in the filename refer to the ID of the specific Huron scanner. Finally, not all sections have been rescanned (6864 out of 7404) and out of these there are 62 duplicate images.

Preview thumbnails of the raw Huron sections at 128 microns. This allows for easy identification and selection of good quality images at 1 micron to be mapped to the 20-micron aligned brain.