Armand1
commented
4 years ago @Yuchen
Where, exactly, did you get the segmented images from that you are working on?
The reason I ask is because the list of segments that you have provided doesn't quite tally with mine. The reason is this: you have a bunch that appear to have been subdivided. For example, where I have:
VA02_0000_t --- both in my images and in my metadata
You have:
VA02_1_0000_t
The extra subscript " 1" comes, I think, from when Salim chopped up images. But I don't have all of those recorded. So that suggests you got the data from Salim.
I am concerned that we are working with two different sets of segmented images. Can you clarify where you got your images from?
I discovered that some of my "motif_id"s in "all_motifs_all_rgb.csv" were wrong. This affected several hundred motifs. I won't go into why, but suffice it to say that I have corrected the error in a new version of the data set that I shall I send you.
In the course of doing so, it also struck me that I had changed the naming system relative to your "motif_props_all_22-05-19_c.csv". What you called "segment_id" I called "motif_id". I did this because you just had a column labelled "ID" and another called "name" which did not seem sufficiently descriptive. But the change risks confusion. So I have semi-retained your labels in the new dataset:
your "segment_id" = my "segment_id" (e.g., AM01_0004_c) your "ID" ="processed_segment_id" (e.g., AM01_0) your "name"=my "image_id" (e.g., AM01)