histolab

Name: histolab
Author: K-Dense-AI

31kstars3.1kforks

How to install histolab?

k-dense-ai/scientific-agent-skills/histolab

$npx -y skills add K-Dense-AI/scientific-agent-skills --skill histolab

Installs into the current project.

Summary

Lightweight WSI tile extraction and preprocessing. Use for basic slide processing, tissue detection, tile extraction, and stain normalization for H&E images. Best for simple pipelines, dataset preparation, and quick tile-based analysis. For advanced spatial proteomics, multiplexe

Files · 1

View on GitHub

SKILL.md

1	# Histolab
2
3	## Overview
4
5	Histolab is a Python library for processing whole slide images (WSI) in digital pathology. It automates tissue detection, extracts informative tiles from gigapixel images, and prepares datasets for deep learning pipelines. The library handles multiple WSI formats, implements sophisticated tissue segmentation, and provides flexible tile extraction strategies.
6
7	## Installation
8
9	Install OpenSlide system libraries first ([OpenSlide download](https://openslide.org/download/)), then install histolab:
10
11	```bash
12	uv pip install histolab
13	```
14
15	For built-in TCGA sample slides via `histolab.data`, also install pooch:
16
17	```bash
18	uv pip install pooch
19	```
20
21	Histolab 0.7.0 (latest stable) supports Python 3.8–3.11 on Linux and macOS. Windows is not supported as of 0.7.0.
22
23	## Quick Start
24
25	Basic workflow for extracting tiles from a whole slide image:
26
27	```python
28	from histolab.slide import Slide
29	from histolab.tiler import RandomTiler
30
31	# Load slide
32	slide = Slide("slide.svs", processed_path="output/")
33
34	# Configure tiler
35	tiler = RandomTiler(
36	tile_size=(512, 512),
37	n_tiles=100,
38	level=0,
39	seed=42
40	)
41
42	# Preview tile locations
43	tiler.locate_tiles(slide, n_tiles=20)
44
45	# Extract tiles
46	tiler.extract(slide)
47	```
48
49	## Core Capabilities
50
51	### 1. Slide Management
52
53	Load, inspect, and work with whole slide images in various formats.
54
55	Common operations:
56	- Loading WSI files (SVS, TIFF, NDPI, etc.)
57	- Accessing slide metadata (dimensions, magnification, properties)
58	- Generating thumbnails for visualization
59	- Working with pyramidal image structures
60	- Extracting regions at specific coordinates
61
62	Key classes: `Slide`
63
64	Reference: `references/slide_management.md` contains comprehensive documentation on:
65	- Slide initialization and configuration
66	- Built-in sample datasets (prostate, ovarian, breast, heart, kidney tissues)
67	- Accessing slide properties and metadata
68	- Thumbnail generation and visualization
69	- Working with pyramid levels
70	- Multi-slide processing workflows
71
72	Example workflow:
73	```python
74	from histolab.slide import Slide
75	from histolab.data import prostate_tissue
76
77	# Load sample data
78	prostate_svs, prostate_path = prostate_tissue()
79
80	# Initialize slide
81	slide = Slide(prostate_path, processed_path="output/")
82
83	# Inspect properties
84	print(f"Dimensions: {slide.dimensions}")
85	print(f"Levels: {slide.levels}")
86	print(f"Magnification: {slide.properties.get('openslide.objective-power')}")
87
88	# Save thumbnail to processed_path
89	from pathlib import Path
90	Path(slide.processed_path).mkdir(parents=True, exist_ok=True)
91	slide.thumbnail.save(Path(slide.processed_path) / f"{slide.name}_thumbnail.png")
92	```
93
94	### 2. Tissue Detection and Masks
95
96	Automatically identify tissue regions and filter background/artifacts.
97
98	Common operations:
99	- Creating binary tissue masks
100	- Detecting largest tissue region
101	- Excluding background and artifacts
102	- Custom tissue segmentation
103	- Removing pen annotations
104
105	Key classes: `TissueMask`, `BiggestTissueBoxMask`, `BinaryMask`
106
107	Reference: `references/tissue_masks.md` contains comprehensive documentation on:
108	- TissueMask: Segments all tissue regions using automated filters
109	- BiggestTissueBoxMask: Returns bounding box of largest tissue region (default)
110	- BinaryMask: Base class for custom mask implementations
111	- Visualizing masks with `locate_mask()`
112	- Creating custom rectangular and annotation-exclusion masks
113	- Mask integration with tile extraction
114	- Best practices and troubleshooting
115
116	Example workflow:
117	```python
118	from histolab.masks import TissueMask, BiggestTissueBoxMask
119
120	# Create tissue mask for all tissue regions
121	tissue_mask = TissueMask()
122
123	# Visualize mask on slide
124	slide.locate_mask(tissue_mask)
125
126	# Get mask array
127	mask_array = tissue_mask(slide)
128
129	# Use largest tissue region (default for most extractors)
130	biggest_mask = BiggestTissueBoxMask()
131	```
132
133	When to use each mask:
134	- `TissueMask`: Multiple tissue sections, comprehensive analysis
135	- `BiggestTissueBoxMask`: Single main tissue section, exclude artifacts (default)
136	- Custom `BinaryMask`: Specific ROI, exclude annotations, custom segmentation
137
138	### 3. Tile Extraction
139
140	Extract smaller regions from large WSI using different strategies.
141
142	Three extraction strategies:
143
144	RandomTiler: Extract fixed number of randomly positioned tiles
145	- Best for: Sampling diverse regions, exploratory analysis, training data
146	- Key parameters: `n_tiles`, `seed` for reprod

People who liked this also viewed

6 picks

Type

histolab

31kstars3.1kforks

View on GitHub

How to install histolab?

k-dense-ai/scientific-agent-skills/histolab

$npx -y skills add K-Dense-AI/scientific-agent-skills --skill histolab

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# Histolab
2
3	## Overview
4
5	Histolab is a Python library for processing whole slide images (WSI) in digital pathology. It automates tissue detection, extracts informative tiles from gigapixel images, and prepares datasets for deep learning pipelines. The library handles multiple WSI formats, implements sophisticated tissue segmentation, and provides flexible tile extraction strategies.
6
7	## Installation
8
9	Install OpenSlide system libraries first ([OpenSlide download](https://openslide.org/download/)), then install histolab:
10
11	```bash
12	uv pip install histolab
13	```
14
15	For built-in TCGA sample slides via `histolab.data`, also install pooch:
16
17	```bash
18	uv pip install pooch
19	```
20
21	Histolab 0.7.0 (latest stable) supports Python 3.8–3.11 on Linux and macOS. Windows is not supported as of 0.7.0.
22
23	## Quick Start
24
25	Basic workflow for extracting tiles from a whole slide image:
26
27	```python
28	from histolab.slide import Slide
29	from histolab.tiler import RandomTiler
30
31	# Load slide
32	slide = Slide("slide.svs", processed_path="output/")
33
34	# Configure tiler
35	tiler = RandomTiler(
36	tile_size=(512, 512),
37	n_tiles=100,
38	level=0,
39	seed=42
40	)
41
42	# Preview tile locations
43	tiler.locate_tiles(slide, n_tiles=20)
44
45	# Extract tiles
46	tiler.extract(slide)
47	```
48
49	## Core Capabilities
50
51	### 1. Slide Management
52
53	Load, inspect, and work with whole slide images in various formats.
54
55	Common operations:
56	- Loading WSI files (SVS, TIFF, NDPI, etc.)
57	- Accessing slide metadata (dimensions, magnification, properties)
58	- Generating thumbnails for visualization
59	- Working with pyramidal image structures
60	- Extracting regions at specific coordinates
61
62	Key classes: `Slide`
63
64	Reference: `references/slide_management.md` contains comprehensive documentation on:
65	- Slide initialization and configuration
66	- Built-in sample datasets (prostate, ovarian, breast, heart, kidney tissues)
67	- Accessing slide properties and metadata
68	- Thumbnail generation and visualization
69	- Working with pyramid levels
70	- Multi-slide processing workflows
71
72	Example workflow:
73	```python
74	from histolab.slide import Slide
75	from histolab.data import prostate_tissue
76
77	# Load sample data
78	prostate_svs, prostate_path = prostate_tissue()
79
80	# Initialize slide
81	slide = Slide(prostate_path, processed_path="output/")
82
83	# Inspect properties
84	print(f"Dimensions: {slide.dimensions}")
85	print(f"Levels: {slide.levels}")
86	print(f"Magnification: {slide.properties.get('openslide.objective-power')}")
87
88	# Save thumbnail to processed_path
89	from pathlib import Path
90	Path(slide.processed_path).mkdir(parents=True, exist_ok=True)
91	slide.thumbnail.save(Path(slide.processed_path) / f"{slide.name}_thumbnail.png")
92	```
93
94	### 2. Tissue Detection and Masks
95
96	Automatically identify tissue regions and filter background/artifacts.
97
98	Common operations:
99	- Creating binary tissue masks
100	- Detecting largest tissue region
101	- Excluding background and artifacts
102	- Custom tissue segmentation
103	- Removing pen annotations
104
105	Key classes: `TissueMask`, `BiggestTissueBoxMask`, `BinaryMask`
106
107	Reference: `references/tissue_masks.md` contains comprehensive documentation on:
108	- TissueMask: Segments all tissue regions using automated filters
109	- BiggestTissueBoxMask: Returns bounding box of largest tissue region (default)
110	- BinaryMask: Base class for custom mask implementations
111	- Visualizing masks with `locate_mask()`
112	- Creating custom rectangular and annotation-exclusion masks
113	- Mask integration with tile extraction
114	- Best practices and troubleshooting
115
116	Example workflow:
117	```python
118	from histolab.masks import TissueMask, BiggestTissueBoxMask
119
120	# Create tissue mask for all tissue regions
121	tissue_mask = TissueMask()
122
123	# Visualize mask on slide
124	slide.locate_mask(tissue_mask)
125
126	# Get mask array
127	mask_array = tissue_mask(slide)
128
129	# Use largest tissue region (default for most extractors)
130	biggest_mask = BiggestTissueBoxMask()
131	```
132
133	When to use each mask:
134	- `TissueMask`: Multiple tissue sections, comprehensive analysis
135	- `BiggestTissueBoxMask`: Single main tissue section, exclude artifacts (default)
136	- Custom `BinaryMask`: Specific ROI, exclude annotations, custom segmentation
137
138	### 3. Tile Extraction
139
140	Extract smaller regions from large WSI using different strategies.
141
142	Three extraction strategies:
143
144	RandomTiler: Extract fixed number of randomly positioned tiles
145	- Best for: Sampling diverse regions, exploratory analysis, training data
146	- Key parameters: `n_tiles`, `seed` for reprod

People who liked this also viewed

6 picks

Type