Search papers, labs, and topics across Lattice.
This paper introduces a decoupled, I/O-aware pipeline for efficiently extracting embeddings from large-scale whole-slide images (WSIs) by separating the workflow into patch generation, parallel embedding inference, and sharded database ingestion. By isolating data movement from computation, the approach minimizes I/O overhead and enhances scalability, allowing for high-throughput processing even in low-resource environments. The results indicate that storage constraints become the primary bottleneck at higher concurrency levels, shifting the perspective on WSI embedding extraction from a compute-centric to a data-centric challenge.
Storage limitations, not compute power, dictate the scalability of whole-slide image embedding extraction, reshaping our approach to handling large-scale medical data.
Whole-slide images (WSIs) are central to computational pathology but are prohibitively large, making patch-based processing the practical unit for foundation model inference. At scale, however, generating and handling massive numbers of patches on quickly introduces significant I/O and orchestration overhead, often dominating end-to-end performance. We present a decoupled, I/O-aware pipeline for large-scale WSI embedding extraction that decomposes the workflow into three stages: (1) patch generation and staging, (2) embarrassingly parallel embedding inference, and (3) sharded vector database ingestion. This design isolates data movement from compute, enabling efficient patch delivery, scalable multi-node inference with minimal communication. The resulting system produces a distributed vector database where embeddings are persistently coupled with rich metadata (e.g., patient, slide, and patch attributes), enabling efficient filtering, retrieval, and downstream reuse. This representation database is compact and reusable for tasks such as retrieval, classification, and few-shot learning, particularly benefiting low-resource environments. We show that decoupling I/O, computation, and ingestion enables high-throughput WSI embedding extraction at scale. By characterizing the scaling envelope, we demonstrate that storage dominates beyond moderate concurrency, reframing WSI embedding extraction as a data-centric systems problem rather than a purely compute-bound workload.
