pdnl-sana 1.0.4

PDNL Semi-Automatic Neuropathology Analyis (SANA)

PDNL Semi Automatic Neuropath Analysis (PDNL-SANA)

About

PDNL-SANA is a python-based package written by the Penn Digital Neuropathology Lab to formalize our methods of IHC quantification. PDNL-SANA includes functions which facilitate extracting pixel data from a Whole Slide Images (WSI), classifying pixels, and converting positive pixel masks to quantifications.

Requirements

python3.9 or greater

Installation

python3 -m pip install pdnl_sana

Getting Started

We provide several example Jupyter notebooks which contain example code blocks utilizing most of SANA's functionality.

• docs/source/examples/example0_prepare_images.ipynb shows how to extract relevant ROI information from a WSI
• docs/source/examples/example1_process_images.ipynb provides a sandbox for the preprocessing and pixel classification methods
• docs/source/examples/example2_normalize_cortex.ipynb illustrates how to deform a curved section of cortex for more optimal quantification
• docs/source/examples/example3_quantification.ipynb has examples of various quantification methods based on the positive pixel masks created by the previous notebooks

For more information, please refer to the Documentation

Roadmap

• GPU Acceleration
• Automatic GM/WM segmentation
• Generic cell detection/segmentation
• Microglia detection/segmentation
• Structure Tensor Analysis