Pathological image compression for big data image analysis: Application to hotspot detection in breast cancer

Area of Science:

• Digital Pathology
• Medical Imaging
• Big Data Analytics

Background:

• Big Data analytics in digital pathology face bottlenecks due to large, high-resolution image datasets.
• Efficient transmission and processing of vast image data are crucial for distributed storage and computing in pathology.

Purpose of the Study:

• To propose a novel pathological image compression framework tailored for Big Data image analysis in digital pathology.
• To minimize data transfer and reduce computational load on decompression engines.

Main Methods:

• The proposed framework utilizes the JPEG2000 Interactive Protocol.
• Integration of the compression framework into hotspot detection for breast biopsy images.

Main Results:

• Significant reduction in data transfer requirements was achieved.
• Considerable decrease in computational demands for image decompression was observed.
• The framework demonstrated efficacy in hotspot detection tasks.

Conclusions:

• The developed pathological image compression framework effectively addresses Big Data challenges in digital pathology.
• The framework offers a viable solution for optimizing storage, transmission, and computation in large-scale digital pathology image analysis.