Goode et al., 2013, OpenSlide: A vendor-neutral software foundation for digital pathology.

# 세줄 요약 #

1. Although widely touted as a replacement for glass slides and microscopes in pathology, digital slides present major challenges in data storage, transmission, processing, and interoperability.
2. In this paper, we present the design and implementation of OpenSlide, a vendor-neutral C library (easily extensible for various programming languages) for reading and manipulating digital slides of diverse vendor formats.
3. OpenSlide is in use today by many academic and industrial organizations worldwide.

 

 

# 상세 리뷰 #

1. Introduction

• Digital pathology will need to offer some benefits (such as improved accuracy, speed, or convenience) to practicing pathologists to encourage them to become more proficient with the new workflows.
• But digital pathology has three technical challenges:
  • 1) the difficulty of processing digital slides using standard image processing tools and libraries because digital slides are enormous relative to ordinary images.
  • 2) each vendor implements its own digital slide formats, libraries, and viewers,
  • 3) few vendors provide libraries and viewers for non-Windows platforms.
• To address these technical challenges, we have created OpenSlide:
  • 1) that is based on the C library, but it is written in other programming languages too (such as Python, Java, ...).
  • 2) OpenSlide currently handles the various vendor formats (DICOM, SVS, MRXS, ...).
  • 3) finally, OpenSlide is supported on Linux, Windows, and on Mac OS X.

OpenSlide Homepage: https://openslide.org/

 

2. Design and Implementation

• OpenSlide design to support various vendor formats, that operates in broad high-level features for vendor formats are shared a common set of problems
  • they pre-compute and store downsampled versions of the full resolution image
  • they are optimized for random access
  • they can store slide metadata and additional small images such as thumbnails
  • they use various kinds of lossy compression
• OpenSlide vendor-neutral API provides read-only access to digital slides stored in a file system.

• The OpenSlide API Provides efficient random access to multiresolution image data using the concept of pyramid levels.
  • level 0 is the highest-resolution level and each subsequent level is a downsampled version of the previous level.
  • The centerpiece is the read_region() function, which extracts a rectangular region of a whole-slide image at a particular pyramid level.
  • All images coordinates are specified with respect to the coordinate system of level 0.

• OpenSlide exposes slide metadata as a set of properties.
• OpenSlide calls relating to associated images
  • small additional images embedded in the slide file
  • such as thumbnails or barcode images.

 

3. Digital Slide Formats

• OpenSlide’s support for many formats has been implemented via empirical analysis of the raw slide data and a significant amount of trial and error.
  • we are often dependent on the OpenSlide user community to contribute digital slides in interesting formats.
• Aperio SVS:
  • Aperio SVS is a single-file
  • TIFF-based format with non-standard metadata (free-form text data) and a specific internal organization.
• 3DHISTECH MRXS (MIRAX):
  • MRXS is a multi-file format (JPEG, PNG, BMP, but OpenSlide support only JPEG) with very complicated metadata in a mixture of text and binary formats,
  • so the MRXS format continues to surprise us and we regularly have to modify OpenSlide to support new variants discovered by the user community.
• Hamamatsu VMS and VMU
  • Hamamatsu VMS is a multi‑file, JPEG‑based format with primarily text‑based metadata
• Leica SCN
  • Leica SCN is a single‑file format based on BigTIFF.
  • Slide metadata is stored in XML format in a TIFF text field.

 

 

* Reference: Goode, Adam, et al. "OpenSlide: A vendor-neutral software foundation for digital pathology." Journal of pathology informatics 4 (2013).