Closed jeffmoffitt closed 2 years ago
I concur. Many of my tissue samples have variable shapes, so using a rectangular grid could image many empty FOVs. I Thought that there were two ways around this: have a simple code that blurs the image composite taken at 10x and then calculates a "mask" using a threshold. This mask would be then used to calculate the tiled positions. The problem with this approach is that the segmentation may vary in quality. Maybe the most error-prone way to get a mask for a arbitrary tissue shape is to draw a polygon over the image, which later can be used to calculate the positions tile.
FWIW I'm working on Steve now. Hopefully once I have Steve cleaned up and better organized it will be easier to add this feature.
I pushed a branch called 'steve_v2'. I tried to clean up and better organize Steve. It should be easier now to add additional modules that support features such as what you requested. Test it first as it still likely has some rough edges..
I just merged 'steve_v2' into master as I'm pretty sure it at least can't be worse than the version of steve in master.
This feature has now been added to the new steve in pull request 123
It would be great if the user could generate a set of tiled positions around different regions of a sample in Steve. Specifically, the user should have control over the number of positions in X and Y, as well as the spacing of these positions. With this control the user could then dynamically adjust the tile to fit different features of the sample.
In parallel, it would be useful if Steve would display these tiles as they are being generated so that the user can confirm that they properly cover the desired features.