medmask 0.1.3

Medical Image Mask Compression and Processing Library

MedMask - Medical Image Mask Processing Library

A specialized library for efficient compression, storage, and processing of medical image segmentation masks. Designed to dramatically improve medical image analysis workflow through advanced compression and semantic design.

🚀 Key Benefits

• Ultimate Compression: 50+ compression ratio using Zstandard
• File Management: 117 files → 1 archive file
• Performance: 16x faster read operations
• Semantic Mapping: Built-in name-to-label conversion
• Overlapping Masks: Multi-granularity organ combinations
• Lazy Loading: Memory-efficient on-demand construction

📦 Installation

pip install medmask

Dependencies: Python 3.8+, numpy, spacetransformer, zstandard

⚡ Usage

Basic Operations

import numpy as np
from medmask import SegmentationMask, MaskArchive, save_mask, load_mask
from spacetransformer import Space

# Create spatial information 
space = Space(shape=(64, 192, 192), spacing=(2.5, 1.0, 1.0))

# -----------------------------
# Create mask – two approaches
# -----------------------------
# 1. Complete initialization
liver_data = np.zeros((64, 192, 192), dtype=np.uint8)  # (Z,Y,X)
liver_data[20:40, 50:150, 60:140] = 1
mask = SegmentationMask(liver_data, {"liver": 1}, space=space)

# 2. Lazy loading (multiple organs)
combined_mask = SegmentationMask.lazy_init(bit_depth=8, space=space)
combined_mask.add_label(liver_data > 0, label=1, name="liver")
combined_mask.add_label(spleen_data > 0, label=2, name="spleen")

# Query masks
liver_region = mask.get_binary_mask_by_names("liver")
multiple_organs = combined_mask.get_binary_mask_by_names(["liver", "spleen"])

File Operations

# Single mask files (.msk)
medmask.save_mask(mask, "liver.msk")
loaded_mask = medmask.load_mask("liver.msk")

# Equivalent
mask.save('liver.msk')
loaded_mask = SegmentationMask.load('liver.msk')

# Multi-mask archives (.mska) - for collections
archive = MaskArchive("organs.mska", mode="w", space=space)
archive.add_segmask(liver_mask, "liver")
archive.add_segmask(heart_mask, "heart")

# Read from archive
archive = MaskArchive("organs.mska", mode="r")
liver = archive.load_segmask("liver")
all_names = archive.all_names()

Use Cases

Single Masks: Individual organ processing, simple storage Combined Masks: Multiple organs in one mask, unified label management
Archives: Collections of related masks, overlapping hierarchies (e.g., individual ribs + combined rib groups)

📊 Performance Comparison

Metric	Traditional	MedMask	Improvement
File Count	117 .nii.gz files	1 .mska file	117:1
Storage Size	5.12 MB	92 KB	56.7:1
Read Time	1.869s	0.117s	16.0x
File Management	Complex	Simple	✓

Test results based on TotalSegmentator 117-organ segmentation data

📁 Project Structure

medmask/
├── core/
│   ├── segmask.py       # SegmentationMask class
│   └── mapping.py       # LabelMapping class
├── storage/
│   ├── archivefile.py   # MaskArchive class
│   └── maskfile.py      # MaskFile class
├── compression/
│   └── zstd_codec.py    # Compression implementation
└── utils/
    └── utils.py         # Utility functions

🧪 Testing

python -m pytest tests/

📝 License

MIT License - see LICENSE file for details.

🤝 Contributing

Issues and Pull Requests are welcome! Please follow the standard GitHub workflow.

---

MedMask - Making medical image segmentation mask processing simpler and more efficient!