meghalithic / magnifica Goto Github PK

There are 19 images with the same box id. We may want to rethink how we label the zooids.

https://pyimagesearch.com/2015/09/07/blur-detection-with-opencv/

Orr et al. 2021:
BLEED 1323

Orr et al. 2022:
BLEED 242
BLEED 80
BLEED 244
BLEED 225
BLEED 686

do not see any asymmetry between right and left height of operculum:

check to see if they have a formation in the metadata file or ask Mali

• waipuru should be same as tewkesbury
• zooids that grew inside an old one
• bump up modern samples

The only difference was that the lab computer has the image "MHR.713.BLEED.582.tif". I am unsure what this image is or of? @klvoje any ideas what the prefix MHR is?

P = triangle = 17
G = square = 15
Global g = diamond

reconcile differences in naming and abbreviation

DATA WRANGLING:

• compare AP metadata file with lab metadata file and images in JPG
• create subset of images to run through Steginator
• add in scale

ANALYSES:

• sum(diag(P)) v sum(diag(G)) (sum of diagonal is like size of the matrix)
• sum(diag(P)) v time
• sum(diag(P)) v change in angle
• conditional evolvability v time, size of matrix, change in angle

Seems most traits have some sort of bimodality in distribution. As an example, see zooid height:

Bimodality is only seen in three formations: Waipurpu, NKBS, Upper Kai-Iwi

Possibly due to image quality? Will need to see boxes and output images that relate to output_21Apr2023_Stegniator.csv.

It seems the first mode ends around 300 pixels, which encompass 95 images (out of 891 images total).

based on #2 , I will change the linear traits image and list on the readme file

modern_metadata: connecting site information with specimen ID. This is now updated in the metadata file for Steginoporella SEMS
modern_locality_metadata: aggregation of metadata for sites used in this study
KLV_metadata: collecting metadata from the images provided by KLV
BLEED_metadata: subset of BLEED image metadata for images used in this study from the BLEED DB

NOTE: NIWA file-name sequence is: Taxon name, NIWA registration number, station number, and image number. A, B, etc, refer to different colonies (or fragments of colonies) of the same species from the same station.

only 10 images removed

From discussion with Carolann:

• in the literature, we'll find traits on maximum width of operculum and maximum height of zooid at the center lines
• suggested that right and left lengths of the cryptocyst may not be that biologically informative
• suggested looking at proportions of b-zooids (both in relative size and in relative number)
• suggested taking base and mid-width measurements
• suggested that maximum height is from landmark 4, not 2
• said that the distance between 20 to 21 is the base of the operculum, not condyles
• suggested also taking a measurement of polypide tube width (between landmarks 5 and 6) as that varies between species
• suggested we don't take averages as there may be asymmetry that is interesting

what does F, S, or P mean?

These come from the "JPG" shared dropbox folder.

There is a discrepancy in two images. One of which is image 313_1.jpg, which AP said to ignore.

found differences in image names between the list of images and the image metadata file, "Imaged Steginoporella magnifica specimens.csv".

Main differences:

• 136 image names in image list that are not in "Imaged Steginoporella magnifica specimens.csv"
• 8 images in "Imaged Steginoporella magnifica specimens.csv" that are not in image list
• found a possible error in an image 237CC
  • has 237_CC_2_15v_x40_BSE replicated
  • has 237_CC_1_15v_x30_BSE_BTK2

BLEED / EPA images:

• 30x: 0.606 pixels per µm
• 35x: 0.705 pixels per µm
• 40x: 0.825 pixels per µm
• 80x: 1.6179 pixels per µm

NIWA images:

• 50x: .404 pixels per µm
• 30x: .242 pixels per µm
• 60x: .484 pixels per µm