 Stephen
|
246fd93525
adam changed
|
2 years ago |
|
Stephen
|
240cf851b9
Changed radom number generator objects to old functions.
|
2 years ago |
|
Stephen
|
9907b2f345
Starting to oldify unet for the supercomputer.
|
2 years ago |
|
Stephen
|
54f869a113
Messed up. Need TPR and FPR. Started making roc curve stuff.
|
2 years ago |
|
Stephen
|
694f95636d
Created helper functions to make main clearer. Created a new conda environment.
|
2 years ago |
|
Stephen
|
21d2def447
Made the confusion mask apply to a single image.
|
2 years ago |
|
Stephen
|
3ee36b9901
Increased the color intensity of the color masks.
|
2 years ago |
|
Stephen
|
bf3ed43398
Successfullyy created the false positive mask.
|
2 years ago |
|
Stephen
|
76c79266c7
COmmented out the reza unet sections.
|
2 years ago |
|
Stephen
|
a470d5c239
Added reza azad's unet file. It doesn't work yet though. It expects inputs of only int32 but my net works on float64.
|
2 years ago |
|
Stephen
|
ad00dbd741
Added the reza azad unet for later testing. Need to change its input size. Currently it takes only grascale images as input at size 256x256.
|
2 years ago |
|
Stephen
|
f5c20f6a41
Found bug in standard unet. 16 was in place of a variable.
|
2 years ago |
|
Stephen
|
88bbedfd7c
Switched to jaccard accuracy score for the monitor. Also saved originals with the output images.
|
2 years ago |
|
Stephen
|
8d24d9184b
Added some printing if statements to clean things up. Added some metrics to the model.
|
2 years ago |
|
Stephen
|
116af0efce
This version makes the correctly predicted white pixels get a red mask image saved to file. Good results start at 3 epochs in terms of redness.
|
3 years ago |
|
Stephen
|
ad4a3d2608
I made a hack to fix the boundary box sizes. The PILLOW function image.paste(im, sizewhatever) seems to be bugged? Or maybe my pictures are strange in some way. To be clear, occasionally image.getbbox() returns something other than (0, 0, width, height), (width, height) come from image.size.
|
3 years ago |
|
Stephen
|
a63b55f1ef
Currently getting a weird error where the image.getbbox() function returns a different size than image.size; BUT ONLY FOR INVERTED IMAGES!
|
3 years ago |
|
Stephen
|
551469d2d8
Ok, the Autoinit thing was messing things up. Working again.
|
3 years ago |
|
Stephen
|
e87dda9966
Something is very wrong. The program no longer creates good outputs.
|
3 years ago |
|
Stephen
|
06e0d9860f
Commented out autoinit for now. About to start going ham on this code.
|
3 years ago |
|
Stephen
|
a52787b597
Added note to remind myself to change the background setting after I fix the foreground/background issue.
|
3 years ago |
|
Stephen
|
81189b7f7a
Updated TODO.
|
3 years ago |
|
Stephen
|
61611d84f4
Updated TODO list.
|
3 years ago |
|
Stephen
|
c33ab62fac
Added the AutoInit thingy. Tested to see it work. Must do tests to see what effect it has.
|
3 years ago |
|
Stephen
|
007429a02e
Implemented a function to create a mask of the prediction with the truth-- Areas in both are shown. Also, I discovered that my model is technically finding the background instead of the foreground. It shouldn't matter much technically, but it explains some of the weird issues I've been having.
|
3 years ago |
|
Stephen
|
80e5619d56
Remembered to use the clever way of multiplication elementwise.
|
3 years ago |
|
Stephen
|
dabe7d428f
Created mask function to output a mask of the whole predicted images.
|
3 years ago |
|
Stephen
|
d6745445cf
update todo
|
3 years ago |
|
Stephen
|
2952a1b013
Added jaccard and dice accuracy score calculations for the predicted full sized images.
|
3 years ago |
|
Stephen
|
2398085d44
Found the error. There was a high in the wide loop. line 595. lol. This version works to produce binarized images of any size!
|
3 years ago |
