Mosaic-Based Color-Transform Optimization for Lossy and Lossy-to-Lossless Compression of Pathology Whole-Slide Images

Area of Science:

• Digital Pathology
• Image Processing
• Data Compression

Background:

• Whole-slide images (WSIs) in pathology are large, requiring efficient storage and transmission.
• Image compression, particularly color transforms, is crucial for managing WSI data.
• Current color transforms do not fully optimize for subsequent compression algorithms, limiting performance.

Purpose of the Study:

• To develop an optimization method for designing irreversible and reversible color transforms for WSIs.
• To ensure color transforms are optimized for specific WSIs and compression algorithms.
• To achieve competitive compression performance with reasonable computational complexity.

Main Methods:

• Proposed mosaic optimization for designing color transforms.
• Simultaneously optimized irreversible and reversible transforms for WSIs and compression algorithms.
• Evaluated performance using JPEG 2000 compression.

Main Results:

• Optimized transforms yielded images more similar to originals compared to state-of-the-art methods.
• Irreversible optimized transforms outperformed Karhunen-Loève Transform in PSNR and visual distortion.
• Optimized transforms improved accuracy in computer-aided nuclei detection tasks.

Conclusions:

• Mosaic optimization provides superior irreversible and reversible color transforms for WSI compression.
• The method enhances image fidelity and diagnostic accuracy in digital pathology.
• Achieved significant improvements in compression metrics and nuclei detection accuracy.