PixelDeck: A local-first media library manager for biomedical imaging

Area of Science:

• Biomedical Imaging
• Computational Pathology
• Digital Pathology

Background:

• Biomedical imaging workflows generate vast amounts of derived image and video data.
• Current methods for managing these assets are often inefficient due to dispersed storage.
• Challenges include retrieval, deduplication, and figure assembly for large media collections.

Purpose of the Study:

• To introduce PixelDeck, an open-source, local-first browser application for organizing and interactively browsing biomedical image and video libraries.
• To provide an efficient and scalable solution for managing high-volume, heterogeneous imaging collections on standard hardware.
• To streamline dataset exploration and preparation for downstream analysis in biomedical research.

Main Methods:

• PixelDeck utilizes a modular ingestion pipeline with recursive folder import and SHA-256-based duplicate detection.
• It features metadata extraction, thumbnail/preview generation, and full-text search within a responsive interface.
• The system employs a Next.js/React frontend, SQLite metadata store (via Prisma), managed local storage, and asynchronous background workers for scalable processing.

Main Results:

• Benchmark imports using public histopathology datasets (PanopTILs, SICAPv2, PanNuke) demonstrated reproducible import behavior and ingestion metrics.
• Duplicate detection and metadata extraction were efficiently handled.
• Embedding-based analysis confirmed dataset-level separation based on image characteristics, validating the system's organization capabilities.

Conclusions:

• PixelDeck offers an efficient and scalable local curation layer for biomedical imaging collections.
• The application facilitates streamlined dataset exploration and preparation for downstream analysis.
• It addresses the challenges of managing dispersed and large-volume imaging assets in modern research workflows.