Safe "cloudification" of large images through picker APIs.

Area of Science:

• Computer Science
• Digital Pathology
• Cloud Computing

Background:

• Generic cloud computing resources are accessible via "drop-file-picker API" (picker API).
• Major cloud platforms (Box.com, Google Drive, Amazon S3) feature picker APIs.
• Whole slide imaging (WSI) in Digital Pathology generates large image files, posing transfer and traversal challenges.

Purpose of the Study:

• To develop a prototype web service application for cloud-based image analysis.
• To expose a "cloudified" API tailored for image analysis using picker APIs.
• To ensure data privacy and governance are maintained during cloud-based analysis.

Main Methods:

• Measured file transfer and traversal times for large WSI image files across cloud resources.
• Developed a prototype web service utilizing picker APIs.
• Created a REST API for image traversal within user-controlled cloud storage.

Main Results:

• Verified extensive interconnectivity between cloud resources.
• Demonstrated a prototype software application exposing an image-traversing REST API.
• Confirmed that image files can be synchronized to cloud resources via picker APIs.

Conclusions:

• A novel, cloud-native API for image analysis was developed, accessible through standard cloud storage.
• The prototype service provides an HTTP REST API for image analysis while maintaining user data governance.
• The open-source prototype is available for public use and further development.