Typical Workflows

Five end-to-end workflows: exploratory tile extraction, comprehensive grid extraction, quality-driven tile selection, multi-slide processing pipelines, and custom tissue detection and filtering.

Typical Workflows

Workflow 1: Exploratory Tile Extraction

Quick sampling of diverse tissue regions for initial analysis.

python

from histolab.slide import Slide
from histolab.tiler import RandomTiler
from pathlib import Path
import logging

# Enable logging for progress tracking
logging.basicConfig(level=logging.INFO)

# Load slide
slide = Slide("slide.svs", processed_path="output/random_tiles/")

# Inspect slide
print(f"Dimensions: {slide.dimensions}")
print(f"Levels: {slide.levels}")
Path(slide.processed_path).mkdir(parents=True, exist_ok=True)
slide.thumbnail.save(Path(slide.processed_path) / f"{slide.name}_thumbnail.png")

# Configure random tiler
random_tiler = RandomTiler(
    tile_size=(512, 512),
    n_tiles=100,
    level=0,
    seed=42,
    check_tissue=True,
    tissue_percent=80.0
)

# Preview locations
random_tiler.locate_tiles(slide, n_tiles=20)

# Extract tiles
random_tiler.extract(slide)

Workflow 2: Comprehensive Grid Extraction

Complete tissue coverage for whole-slide analysis.

python

from histolab.slide import Slide
from histolab.tiler import GridTiler
from histolab.masks import TissueMask

# Load slide
slide = Slide("slide.svs", processed_path="output/grid_tiles/")

# Use TissueMask for all tissue sections
tissue_mask = TissueMask()
slide.locate_mask(tissue_mask)

# Configure grid tiler
grid_tiler = GridTiler(
    tile_size=(512, 512),
    level=1,  # Use level 1 for faster extraction
    pixel_overlap=0,
    check_tissue=True,
    tissue_percent=70.0
)

# Preview grid
grid_tiler.locate_tiles(slide)

# Extract all tiles
grid_tiler.extract(slide, extraction_mask=tissue_mask)

Workflow 3: Quality-Driven Tile Selection

Extract most informative tiles based on nuclei density.

python

from histolab.slide import Slide
from histolab.tiler import ScoreTiler
from histolab.scorer import NucleiScorer
import pandas as pd
import matplotlib.pyplot as plt

# Load slide
slide = Slide("slide.svs", processed_path="output/scored_tiles/")

# Configure score tiler
score_tiler = ScoreTiler(
    tile_size=(512, 512),
    n_tiles=50,
    level=0,
    scorer=NucleiScorer(),
    check_tissue=True
)

# Preview top tiles
score_tiler.locate_tiles(slide, n_tiles=15)

# Extract with report
score_tiler.extract(slide, report_path="tiles_report.csv")

# Analyze scores
report_df = pd.read_csv("tiles_report.csv")
plt.hist(report_df['score'], bins=20, edgecolor='black')
plt.xlabel('Tile Score')
plt.ylabel('Frequency')
plt.title('Distribution of Tile Scores')
plt.show()

Workflow 4: Multi-Slide Processing Pipeline

Process entire slide collection with consistent parameters.

python

from pathlib import Path
from histolab.slide import Slide
from histolab.tiler import RandomTiler
import logging

logging.basicConfig(level=logging.INFO)

# Configure tiler once
tiler = RandomTiler(
    tile_size=(512, 512),
    n_tiles=50,
    level=0,
    seed=42,
    check_tissue=True
)

# Process all slides
slide_dir = Path("slides/")
output_base = Path("output/")

for slide_path in slide_dir.glob("*.svs"):
    print(f"\nProcessing: {slide_path.name}")

    # Create slide-specific output directory
    output_dir = output_base / slide_path.stem
    output_dir.mkdir(parents=True, exist_ok=True)

    # Load and process slide
    slide = Slide(slide_path, processed_path=output_dir)

    # Save thumbnail for review
    Path(slide.processed_path).mkdir(parents=True, exist_ok=True)
    slide.thumbnail.save(Path(slide.processed_path) / f"{slide.name}_thumbnail.png")

    # Extract tiles
    tiler.extract(slide)

    print(f"Completed: {slide_path.name}")

Workflow 5: Custom Tissue Detection and Filtering

Handle slides with artifacts, annotations, or unusual staining.

python

from histolab.slide import Slide
from histolab.masks import TissueMask
from histolab.tiler import RandomTiler
from histolab.filters.compositions import Compose
from histolab.filters.image_filters import RgbToGrayscale, OtsuThreshold
from histolab.filters.morphological_filters import (
    BinaryDilation, RemoveSmallObjects, RemoveSmallHoles
)

# Define custom filter pipeline for aggressive artifact removal
aggressive_filters = Compose([
    RgbToGrayscale(),
    OtsuThreshold(),
    BinaryDilation(disk_size=10),
    RemoveSmallHoles(area_threshold=5000),
    RemoveSmallObjects(area_threshold=3000)  # Remove larger artifacts
])

# Create custom mask
custom_mask = TissueMask(filters=aggressive_filters)

# Load slide and visualize mask
slide = Slide("slide.svs", processed_path="output/")
slide.locate_mask(custom_mask)

# Extract with custom mask
tiler = RandomTiler(tile_size=(512, 512), n_tiles=100)
tiler.extract(slide, extraction_mask=custom_mask)