Content-based Histopathology Image Retrieval using a CometCloud-based infrastructure
Abstract
Research in content-based image retrieval (CBIR) has emerged as an important focus of investigation multiple image-related disciplines. We explore a broad spectrum of potential clinical applications in pathology with a newly developed set of retrieval algorithms that were fine-tuned for each class of digital pathology images, including peripheral blood smears, mammary glands, glomeruli (kidney) and Hematoxylin-stained breast tissue microarray (TMA). We utilized the CometCloud autonomic cloud engine to run the CBIR algorithms in parallel across federated resources.
Intellectual Merit
The goal of this research is to develop algorithms and evaluate them on multiple histopathology image datasets. Moreover, a computational framework will be developed to orchestrate the workflow of the algorithms and the computational resources.
Broader Impact
The combination of this techniques and the computational framework will allow doctors to quickly detect possible pathologies in patients.
Use of FutureGrid
We intend to use the FutureGrid resources to compute our the scientific problem as part of our federation.
Scale Of Use
We will mainly use VMs to run our experiments