Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Processing resolution #2

Merged
merged 10 commits into from
Feb 8, 2024
Merged

Processing resolution #2

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
10 commits
Select commit Hold shift + click to select a range
d624149
udpated to new TPTBox api
Hendrik-code Feb 6, 2024
9be3a19
updated to new TPTBox api
Hendrik-code Feb 6, 2024
8098142
renaming to calm ruff
Hendrik-code Feb 6, 2024
a93c36b
ruff
Hendrik-code Feb 6, 2024
5c44ffe
made spline_param lower to be faster
Hendrik-code Feb 6, 2024
cd9675b
added fix_vertebra_inconsistency processing, some ruff
Hendrik-code Feb 6, 2024
0fdd8fa
added preprocessing as phase
Hendrik-code Feb 7, 2024
eba47b4
framework now saves raw in modelresolution and resamples back to inpu…
Hendrik-code Feb 7, 2024
571c501
adapted to new np_utils TPTBox
Hendrik-code Feb 8, 2024
20c0099
another TPTBox update
Hendrik-code Feb 8, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
67 changes: 29 additions & 38 deletions spineps/phase_instance.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,7 +1,7 @@
# from utils.predictor import nnUNetPredictor
import numpy as np
from TPTBox import NII, Location, Log_Type
from TPTBox.core.np_utils import np_calc_crop_around_centerpoint, np_dice
from TPTBox.core.np_utils import np_calc_crop_around_centerpoint, np_count_nonzero, np_dice, np_unique
from tqdm import tqdm

from spineps.seg_enums import ErrCode, OutputType
Expand All @@ -14,7 +14,7 @@ def predict_instance_mask(
seg_nii: NII,
model: Segmentation_Model,
debug_data: dict,
pad_size: int = 2,
pad_size: int = 0,
fill_holes: bool = True,
use_height_estimate: bool = False,
proc_corpus_clean: bool = True,
Expand Down Expand Up @@ -59,7 +59,8 @@ def predict_instance_mask(
zms = seg_nii_rdy.zoom
logger.print("zms", zms, verbose=verbose)
expected_zms = model.calc_recommended_resampling_zoom(seg_nii_rdy.zoom)
seg_nii_rdy.rescale_(expected_zms, verbose=logger) # in PIR
if not seg_nii_rdy.assert_affine(zoom=expected_zms, raise_error=False):
seg_nii_rdy.rescale_(expected_zms, verbose=logger) # in PIR
#
seg_nii_uncropped = seg_nii_rdy.copy()
logger.print(
Expand All @@ -69,9 +70,9 @@ def predict_instance_mask(
uncropped_vert_mask = np.zeros(seg_nii_uncropped.shape)
logger.print("Vertebra uncropped_vert_mask empty", uncropped_vert_mask.shape, verbose=verbose)
#
crop = seg_nii_rdy.compute_crop_slice(dist=5)
crop = seg_nii_rdy.compute_crop(dist=5)
# logger.print("Crop", crop, verbose=verbose)
seg_nii_rdy.apply_crop_slice_(crop)
seg_nii_rdy.apply_crop_(crop)
logger.print(f"Crop down from {uncropped_vert_mask.shape} to {seg_nii_rdy.shape}", verbose=verbose)
# arr[crop] = X, then set nifty to arr
logger.print("Vertebra seg_nii_rdy", seg_nii_rdy.zoom, seg_nii_rdy.orientation, seg_nii_rdy.shape, verbose=verbose)
Expand All @@ -83,7 +84,9 @@ def predict_instance_mask(
vert_size_threshold = max(int(vert_size_threshold / (expected_zms[0] * expected_zms[1] * expected_zms[2])), 40)

seg_labels = seg_nii.unique()
assert 49 in seg_labels, f"no corpus ({Location.Vertebra_Corpus_border.value}) labels in this segmentation, cannot proceed"
if 49 not in seg_labels:
logger.print(f"no corpus ({Location.Vertebra_Corpus_border.value}) labels in this segmentation, cannot proceed", Log_Type.FAIL)
return None, ErrCode.EMPTY

# get all the 3vert predictions
vert_predictions, hierarchical_existing_predictions, n_corpus_coms = collect_vertebra_predictions(
Expand Down Expand Up @@ -144,17 +147,17 @@ def predict_instance_mask(
debug_data["inst_uncropped_vert_arr_a"] = whole_vert_nii_uncropped.copy()

# Resample back to input space
whole_vert_nii_uncropped.rescale_(zms, verbose=verbose)
debug_data["inst_uncropped_vert_arr_b_rescale"] = whole_vert_nii_uncropped.copy()
whole_vert_nii_uncropped.reorient_(orientation, verbose=verbose)
debug_data["inst_uncropped_vert_arr_c_reorient"] = whole_vert_nii_uncropped.copy()
# whole_vert_nii_uncropped.rescale_(zms, verbose=verbose)
# debug_data["inst_uncropped_vert_arr_b_rescale"] = whole_vert_nii_uncropped.copy()
# whole_vert_nii_uncropped.reorient_(orientation, verbose=verbose)
# debug_data["inst_uncropped_vert_arr_c_reorient"] = whole_vert_nii_uncropped.copy()
if pad_size > 0:
# logger.print(whole_vert_nii_uncropped.shape)
arr = whole_vert_nii_uncropped.get_array()
arr = arr[pad_size:-pad_size, pad_size:-pad_size, pad_size:-pad_size]
whole_vert_nii_uncropped.set_array_(arr)
# logger.print(whole_vert_nii_uncropped.shape)
whole_vert_nii_uncropped.pad_to(shp, inplace=True)
# whole_vert_nii_uncropped.pad_to(shp, inplace=True)

return whole_vert_nii_uncropped, ErrCode.OK

Expand Down Expand Up @@ -197,6 +200,10 @@ def collect_vertebra_predictions(
corpus_coms.reverse() # from bottom to top
n_corpus_coms = len(corpus_coms)

if n_corpus_coms < 3:
logger.print(f"Too few vertebra semantically segmented ({n_corpus_coms})", Log_Type.FAIL)
return None, [], 0

shp = (
# n_corpus_coms,
# 3
Expand Down Expand Up @@ -248,26 +255,7 @@ def collect_vertebra_predictions(
break
seg_at_com = seg_arr_c[int(com[0])][int(com[1])][int(com[2])] != 0

if use_height_estimate:
com_above = 0 if com_idx < 2 else com_idx - 2
com_above = corpus_coms[com_above] # (com, bbox)
com_below = n_corpus_coms - 1 if com_idx > n_corpus_coms - 3 else com_idx + 2
com_below = corpus_coms[com_below]
com_y = com[1]
height = min(
int(max(abs(com_above[1] - com_y), abs(com_below[1] - com_y)) * 2.5), # type:ignore
cutout_size[1],
)
if len(corpus_coms) <= 3:
height = min(int(abs(com_above[1] - com_below[1]) * 2.5), cutout_size[1])
height = height + 1 if height % 2 != 0 else height
cutout_size2 = (
cutout_size[0],
height,
cutout_size[2],
)
else:
cutout_size2 = cutout_size
cutout_size2 = cutout_size

# Calc cutout
arr_cut, cutout_coords, paddings = np_calc_crop_around_centerpoint(com, seg_arr_c, cutout_size2)
Expand All @@ -278,15 +266,18 @@ def collect_vertebra_predictions(
cut_nii,
resample_to_recommended=False,
pad_size=0,
# resample_output_to_input_space=False,
resample_output_to_input_space=False,
verbose=False,
)
vert_cut_nii = results[OutputType.seg_modelres].reorient_()
vert_cut_nii = results[OutputType.seg].reorient_()
# print("vert_cut_nii", vert_cut_nii.shape)
# logger.print(f"Done {com_idx}")
debug_data[f"inst_cutout_vert_nii_{com_idx}_pred"] = vert_cut_nii.copy()
vert_cut_nii = post_process_single_3vert_prediction(
vert_cut_nii, None, fill_holes=fill_holes, largest_cc=proc_largest_cc # type:ignore
vert_cut_nii,
None,
fill_holes=fill_holes,
largest_cc=proc_largest_cc, # type:ignore
)
vert_labels = vert_cut_nii.unique() # 1,2,3
debug_data[f"inst_cutout_vert_nii_{com_idx}_proc"] = vert_cut_nii.copy()
Expand Down Expand Up @@ -527,13 +518,13 @@ def merge_coupled_predictions(
combine[combine < m] = 0
combine[combine != 0] = idx

count_new = np.count_nonzero(combine)
count_new = np_count_nonzero(combine)
if count_new == 0:
logger.print("ZERO instance mask failure on vertebra instance creation", Log_Type.FAIL)
return seg_nii, debug_data, ErrCode.EMPTY
fixed_n = combine.copy()
fixed_n[whole_vert_arr != 0] = 0
count_cut = np.count_nonzero(fixed_n)
count_cut = np_count_nonzero(fixed_n)
relative_overlap = (count_new - count_cut) / count_new
if relative_overlap > 0.6:
logger.print(k, f" was skipped because it overlaps {round(relative_overlap, 4)} with established verts", verbose=verbose)
Expand All @@ -543,15 +534,15 @@ def merge_coupled_predictions(

debug_data["inst_crop_vert_arr_a_raw"] = seg_nii.set_array(whole_vert_arr)

if len(np.unique(whole_vert_arr)) == 1:
if len(np_unique(whole_vert_arr)) == 1:
logger.print("Vert mask empty, will skip", Log_Type.FAIL)
return whole_vert_nii.set_array_(whole_vert_arr, verbose=False), debug_data, ErrCode.EMPTY

# Cleanup step
if proc_cleanvert:
whole_vert_arr = clean_cc_artifacts(
whole_vert_arr,
labels=np.unique(whole_vert_arr)[1:], # type:ignore
labels=np_unique(whole_vert_arr)[1:], # type:ignore
cc_size_threshold=vert_size_threshold,
only_delete=True,
logger=logger,
Expand Down
Loading
Loading