Closed berithi closed 6 years ago
I found the problem. My autohistogram function fails if the number of bins in the histogram exceeds 10 billion or so. :) There were some large outliers. I now modified the function so that if it would be too large, it trims the very large values off the edges.
This requires an update to vhlab_mltbx_toolbox, there's actually nothing wrong with vhlab_mlapp_ArrayTomography in this regard.
However, the ROIs in this vglut1 dataset don't look good, and i think it is because the thresholding is not done well. There seems to be a lot of background points, unless I am not understanding what the image is very well.
Thanks! Yes- I was struggling with the thresholding (it included all the nuclei) and I was trying to see if I could get around that problem with the volume filtering... b.
On Sat, Jan 20, 2018 at 9:37 PM, Steve Van Hooser notifications@github.com wrote:
I found the problem. My autohistogram function fails if the number of bins in the histogram exceeds 10 billion or so. :) There were some large outliers. I now modified the function so that if it would be too large, it trims the very large values off the edges.
This requires an update to vhlab_mltbx_toolbox, there's actually nothing wrong with vhlab_mlapp_ArrayTomography in this regard.
However, the ROIs in this vglut1 dataset don't look good, and i think it is because the thresholding is not done well. There seems to be a lot of background points, unless I am not understanding what the image is very well.
— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/VH-Lab/vhlab_mlapp_ArrayTomography/issues/8#issuecomment-359219180, or mute the thread https://github.com/notifications/unsubscribe-auth/AeniDBBnZOmHp4fypzMlgjX46zkcVDq3ks5tMqLagaJpZM4RlDzd .
-- Berith Isaac berithi@gmail.com 617-301-0251
This happens only for the vglut1 dataset in YEL22JAN_1_1... Thanks :-) b.