HistomicsTK

1.4.0 verified Fri May 01 auth: no python

HistomicsTK is a Python toolkit for histopathology image analysis, providing algorithms for segmentation, feature extraction, and classification of whole-slide images. Current version is 1.4.0, with a release cadence of approximately 2-3 months.

pip install histomicstk

Common errors

error ModuleNotFoundError: No module named 'histomicstk' ↓

cause Package not installed or incorrect environment.

fix

Run: pip install histomicstk

error ModuleNotFoundError: No module named 'large_image' ↓

cause Optional dependency not installed for whole-slide image support.

fix

Run: pip install large-image or pip install histomicstk[all]

error ValueError: Haralick features require a 2-D image ↓

cause Input image has more than 2 dimensions (e.g., RGB).

fix

Convert to grayscale: from skimage.color import rgb2gray; gray = rgb2gray(image)

error AttributeError: module 'histomicstk.features' has no attribute 'compute_haralick_features' ↓

cause Old import path used or wrong version.

fix

Ensure you have v1.4.0+ and use correct import: from histomicstk.features import compute_haralick_features

Warnings

breaking In v1.4.0, Python 3.8 support was dropped. You must use Python >=3.9. ↓

fix Upgrade Python to 3.9+ or pin histomicstk to <1.4.0.

deprecated Direct import of submodules like histomicstk.features.haralick is deprecated. Use histomicstk.features instead. ↓

fix Update imports to use top-level module: from histomicstk.features import compute_haralick_features

gotcha compute_haralick_features expects float32 image in [0,1] range. Passing integer or unnormalized images yields incorrect results. ↓

fix Normalize image: image = image.astype(np.float32) / 255.0

gotcha Nuclear segmentation functions require stain normalization (e.g., Reinhard) before use. Skipping this step leads to poor segmentation. ↓

fix Use histomicstk.preprocessing.color_normalization.reinhard first.

Install

pip install histomicstk[all]

Imports

histomicstk
wrong
```
import HistomicsTK
```
correct
```
import histomicstk as htk
```
Case-sensitive; package is lowercase

segmentation.nuclear_segmentation

wrong

from histomicstk.segmentation.nuclei import nuclear_segmentation

correct

from histomicstk.segmentation import nuclear_segmentation

Old path deprecated; use nuclear_segmentation submodule

features.compute_haralick_features

wrong

from histomicstk.features.haralick import compute_haralick_features

correct

from histomicstk.features import compute_haralick_features

Haralick features moved to top-level features module in v1.4.0

Quickstart

Basic example of computing Haralick features from a 2D image. For whole-slide images, you need to install 'large_image' and use it to read tiles.

import histomicstk as htk
import numpy as np

# Load a sample image (replace with your own)
# For whole-slide images, use large_image
# Here we simulate a 2D grayscale image
image = np.random.rand(100, 100).astype(np.float32)

# Compute Haralick texture features
from histomicstk.features import compute_haralick_features
haralick = compute_haralick_features(image, return_labels=True)
print(haralick['Harmonic mean'])

# Nuclear segmentation (requires stain normalization first)
from histomicstk.segmentation import nuclear_segmentation
# nuclei_df, nuclei_im = nuclear_segmentation.nuclei_segmentation(image)  # requires proper staining