SPECIAL NOTICE
99 -- Advances in Artificial Intelligence to Generate AIMI Annotations
- Notice Date
- 1/24/2023 10:46:44 AM
- Notice Type
- Special Notice
- NAICS
- 541714
— Research and Development in Biotechnology (except Nanobiotechnology)
- Contracting Office
- NIH NCI Bethesda MD 20892 USA
- ZIP Code
- 20892
- Solicitation Number
- S23-040
- Response Due
- 3/10/2023 12:00:00 PM
- Point of Contact
- Jo Brown, Phone: 301-846-5720
- E-Mail Address
-
jo.brown@nih.gov
(jo.brown@nih.gov)
- Description
- We are seeking vendors who can provide support in AIMI Annotations:� Advances in Artificial intelligence (AI) in Medical Imaging (MI) have led to a profusion of studies that apply Deep Learning (DL) models to problems in radiology and pathology. To develop reliable AI models for cancer imaging and potential clinic use, the availability of cancer images with high-quality annotations is crucial. However, manual annotation by experts is not feasible for large-scale datasets because of the level of effort needed by highly trained experts. Over the past several years, there have been significant advancements in the development of models that perform image segmentation of tumors with high accuracy.� Applications of advanced DL models to publicly available imaging data can help generate annotations in the form of tumor segmentations that would be useful to the research community in the downstream development of medical imaging AI models.� The Imaging Data Commons (IDC) connects researchers with publicly available cancer imaging data, often linked with other types of cancer data, and co-located with cloud-based computational resources�and big data analysis tools provided by various cloud platforms, such Google Cloud Platform (GCP). It also provides the tools to search and visualize de-identified cancer imaging data, define cohorts, and use those cohorts for cloud-based analysis to better understand the disease and evaluate treatment options.� Over the past several years, the development of DL-based algorithms has dramatically flourished, due in part to advances in computation hardware and the availability of large datasets. Some DL-based algorithms in medical imaging offer high-quality segmentations, classifications/staging, and regression predictions. The availability of high-quality annotations is of importance to develop cutting-edge AI-based algorithms for cancer imaging applications. Currently, many image collections in IDC lack reliable annotations of tumors, organs, or tissues. The availability of reliable annotations is critical to supervise machine learning-based algorithms, particularly for DL-based algorithms.� Because manual segmentation by experts can be costly and time-consuming, full-/semi- automatic, fast, computer-aided algorithms for generating high-quality annotations are in high demand.� Machine-generated annotations can be further refined/verified by experts. Finally, expert-verified annotations can be used to facilitate relevant research. There are some publicly available DL-based models, such as models in AIMI Hub (http://modelhub.ai/) and MONAI (https://monai.io/). The models must be well-known (such as nnUNet, efficientNet, etc.) or be able to offer state-of-the-art performance. The goal of the AI in Medical Imaging (AIMI) initiative is to solicit proposals for curation of derived datasets in the form of advanced machine-generated segmentation of tumor and organ contents of radiology collections, hereafter referred to as AIMI Annotations in the NCI Imaging Data Commons. For information concerning opportunities with the FNLCR, please contact the Leidos Biomedical Research Point of Contact listed in this announcement. Background Information:� Leidos Biomedical Research, Inc. (Leidos Biomed), a subsidiary of Leidos Corporation, a Delaware corporation with offices in Frederick, MD is the Operations and Technical Support (OTS) contractor for the National Cancer Institute at Frederick.� The effort to be performed will be part of Leidos Biomed�s Prime Contract 75N91019D00024 that has been issued by the National Cancer Institute, National Institutes of Health, Frederick, MD. Leidos Biomedical Research, Inc.� POC:� Lead Subcontracts Administrator Jo Brown at jo.brown@nih.gov Response Dates:� Vendors Request for Proposal � Up to February 13, 2023; RFP will be distributed upon request Questions Due; February 21 ,2023 at 3:00 PM Proposals Due by March 10, 2023, at 3:00 PM
- Web Link
-
SAM.gov Permalink
(https://sam.gov/opp/c8e97bbc54944eda92645e1629df8228/view)
- Record
- SN06571539-F 20230126/230124230104 (samdaily.us)
- Source
-
SAM.gov Link to This Notice
(may not be valid after Archive Date)
| FSG Index | This Issue's Index | Today's SAM Daily Index Page |