Histolab：病理切片的预处理工具｜项目实战

小罗碎碎念

今天这期推文主要介绍一个病理切片的预处理工具：Histolab。

通过这期推文你最终可以实现的效果如下：

这只是一个简单的示意图，你看完本期推文以后可以自行调整。后期还会继续更新其他的模型，由易到难，现在看不上，小心以后跟不上哟，哈哈。

---

视频教程

复制网址，浏览器打开。

https://www.bilibili.com/video/BV1o6s4eoEwY/?vd_source=7c7c08bd51e5e86f3ee3a8766822236b

---

一、安装依赖项

随便打开一个虚拟环境，输入

pip install histolab

---

二、导入数据

2-1：确定路径

from histolab.slide import Slide
import os
BASE_PATH = os.getcwd()

print(BASE_PATH)

---

2-2：设置路径

CRC_path = '/data/data_mailab015/病理AI/人工智能病理/2024病理汇总-WDX/贵肿-svs/Raw-svs/1800518-1.svs'  # location
path = r"./Histo"  # 切割后图片存放的位置
CRC_slide = Slide(CRC_path,processed_path=path)

---

2-3：获取切片信息

print(f"Slide name:{CRC_slide.name}") # 幻灯片名称
print(f"Levels:{CRC_slide.levels}")
print(f"Dimensions at level 0:{CRC_slide.dimensions}")
print(f"Dimensions at level 1:{CRC_slide.level_dimensions(level=1)}")
print(f"Dimensions at level 2:{CRC_slide.level_dimensions(level=2)}")

---

2-4：获取缩略图

CRC_slide.thumbnail

---

三、组织提取

3-1：随机提取

from histolab.tiler import RandomTiler

random_tiles_extractor = RandomTiler(
    tile_size=(224, 224),
    n_tiles=30,
    level=0,
    seed=42,
    check_tissue=True, # default
    tissue_percent=80.0, # default
    prefix="random12/", # save tiles in the "random" subdirectory of slide's processed_path
    suffix=".png" # default
)

random_tiles_extractor.locate_tiles(
    slide=CRC_slide,
    scale_factor=24, # default
    alpha=128, # default
    outline="red", # default
)

提取切片

random_tiles_extractor.extract(CRC_slide)

---

3-2：网格提取

from histolab.tiler import GridTiler

grid_tiles_extractor = GridTiler(
   tile_size=(512, 512),
   level=0,
   check_tissue=False,
   pixel_overlap=0, # default
   prefix="grid/", # save tiles in the "grid" subdirectory of slide's processed_path
   suffix=".png" # default
)

grid_tiles_extractor.locate_tiles(
    slide=CRC_slide,
    scale_factor=64,
    alpha=64,
    outline="#046C4C",
)

grid_tiles_extractor.extract(CRC_slide)

---

3-3：分数提取

from histolab.tiler import ScoreTiler
from histolab.scorer import NucleiScorer

scored_tiles_extractor = ScoreTiler(
    scorer = NucleiScorer(),
    tile_size=(512, 512),
    n_tiles=100,
    level=0,
    check_tissue=True,
    tissue_percent=80.0,
    pixel_overlap=0, # default
    prefix="scored/", # save tiles in the "scored" subdirectory of slide's processed_path
    suffix=".png" # default
)

grid_tiles_extractor.locate_tiles(slide=CRC_slide)