Napari中的粉虱胚胎形态测量随时间变化#
作者：Robert Haase, Daniela Vorkel, 2020
这是之前为clij2发布的工作流程的pyclesperanto版本。 ImageJ宏原版
该脚本是大量GPU加速处理的一个示例。建议使用具有至少8 GB GDDR6内存的专用 显卡。否则，可能会相当慢。
让我们首先检查pyclesperanto是否已安装以及它使用的是哪个GPU。
import pyclesperanto_prototype as cle
import numpy as np
cle.select_device("RTX")
<Apple M1 Max on Platform: Apple (2 refs)>
加载数据集#
该数据集显示了一个粉虱胚胎，由定制的光片显微镜在488nm波长下成像（成像致谢：Daniela Vorkel, Myers实验室, MPI CBG）。可以从zenodo下载。 数据集已重新采样为1x1x1微米的体素大小。胚胎表达细胞核GFP。我们将使用该数据集来检测细胞核并生成估计的细胞分割。
所有处理步骤都在3D空间中执行。
from skimage.io import imread
timelapse = imread('/Users/haase/data/Lund_18.0_22.0_Hours-resampled.tif')
# 打印图像的空间维度
print(timelapse.shape)
(25, 140, 532, 266)
def process_image(image):
import time
start_time = time.time()
# 将图像推送到GPU内存并显示
gpu_input = cle.push(image)
# print(gpu_input)
# 高斯模糊
sigma = 2.0
gpu_blurred = cle.gaussian_blur(gpu_input, sigma_x=sigma, sigma_y=sigma, sigma_z=sigma)
# 检测最大值
gpu_detected_maxima = cle.detect_maxima_box(gpu_blurred)
# 阈值
threshold = 300.0
gpu_thresholded = cle.greater_constant(gpu_blurred, constant=threshold)
# 掩码
gpu_masked_spots = cle.mask(gpu_detected_maxima, gpu_thresholded)
# 标记斑点
gpu_labelled_spots = cle.connected_components_labeling_box(gpu_masked_spots)
# show_labels(gpu_labelled_spots)
number_of_spots = int(cle.maximum_of_all_pixels(gpu_labelled_spots))
# print("检测到的斑点数量: " + str(number_of_spots))
# 标签图闭合
number_of_dilations = 10
flip = cle.create_labels_like(gpu_labelled_spots)
flop = cle.create_labels_like(gpu_labelled_spots)
flag = cle.create([1,1,1])
cle.copy(gpu_labelled_spots, flip)
for i in range (0, number_of_dilations) :
cle.onlyzero_overwrite_maximum_box(flip, flag, flop)
cle.onlyzero_overwrite_maximum_diamond(flop, flag, flip)
# 腐蚀标签
flap = cle.greater_constant(flip, constant=1)
number_of_erosions = 4
for i in range(0, number_of_erosions):
cle.erode_box(flap, flop)
cle.erode_box(flop, flap)
gpu_labels = cle.mask(flip, flap)
# 将结果从GPU作为numpy数组获取回来
result = cle.pull(gpu_labels).astype(np.uint16)
print("处理耗时 " + str(time.time() - start_time) + " 秒")
return result
from skimage import data
import napari
viewer = napari.Viewer()
Warning: Could not find scikit-tensor which is needed for separable approximations...
If you want to compute separable approximations, please install it with
pip install scikit-tensor-py3
# 体素大小 z,y,x
calibration = [1, 1, 1]
# 用于可视化的便捷函数
def show(image):
viewer.add_image(image, scale=calibration)
def show_labels(labels):
viewer.add_labels(labels, scale=calibration)
# 改编自：https://github.com/tlambert03/napari-dask-example/blob/master/dask_napari.ipynb
import dask
import dask.array as da
# 创建延迟图像读取器的dask堆栈
lazy_process_image = dask.delayed(process_image) # 延迟读取器
lazy_arrays = [lazy_process_image(timelapse[n]) for n in range(0, timelapse.shape[0])]
dask_arrays = [
da.from_delayed(lazy_array, shape=timelapse[0].shape, dtype=np.uint16)
for lazy_array in lazy_arrays
]
# 堆叠成一个大的dask.array
dask_stack = da.stack(dask_arrays, axis=0)
dask_stack
|
print(timelapse.shape[0])
25
show(timelapse)
show_labels(dask_stack)
Processing took 0.3502938747406006 s
Processing took 0.268435001373291 s
Processing took 0.2793867588043213 s
from napari.utils import nbscreenshot
nbscreenshot(viewer)
