Showing 1–4 of 4 results for author: Barnowski, R

A Foundation Model for Cell Segmentation

Authors: Uriah Israel, Markus Marks, Rohit Dilip, Qilin Li, Morgan Schwartz, Elora Pradhan, Edward Pao, Shenyi Li, Alexander Pearson-Goulart, Pietro Perona, Georgia Gkioxari, Ross Barnowski, Yisong Yue, David Van Valen

Abstract: Cells are the fundamental unit of biological organization, and identifying them in imaging data - cell segmentation - is a critical task for various cellular imaging experiments. While deep learning methods have led to substantial progress on this problem, models that have seen wide use are specialist models that work well for specific domains. Methods that have learned the general notion of "what… ▽ More Cells are the fundamental unit of biological organization, and identifying them in imaging data - cell segmentation - is a critical task for various cellular imaging experiments. While deep learning methods have led to substantial progress on this problem, models that have seen wide use are specialist models that work well for specific domains. Methods that have learned the general notion of "what is a cell" and can identify them across different domains of cellular imaging data have proven elusive. In this work, we present CellSAM, a foundation model for cell segmentation that generalizes across diverse cellular imaging data. CellSAM builds on top of the Segment Anything Model (SAM) by develo** a prompt engineering approach to mask generation. We train an object detector, CellFinder, to automatically detect cells and prompt SAM to generate segmentations. We show that this approach allows a single model to achieve state-of-the-art performance for segmenting images of mammalian cells (in tissues and cell culture), yeast, and bacteria collected with various imaging modalities. To enable accessibility, we integrate CellSAM into DeepCell Label to further accelerate human-in-the-loop labeling strategies for cellular imaging data. A deployed version of CellSAM is available at https://label-dev.deepcell.org/. △ Less

Submitted 18 November, 2023; originally announced November 2023.

3D Gamma-ray and Neutron Map** in Real-Time with the Localization and Map** Platform from Unmanned Aerial Systems and Man-Portable Configurations

Authors: R. Pavlovsky, J. W. Cates, W. J. Vanderlip, T. H. Y. Joshi, A. Haefner, E. Suzuki, R. Barnowski, V. Negut, A. Moran, K. Vetter, B. J. Quiter

Abstract: Nuclear Scene Data Fusion (SDF), implemented in the Localization and Map** Platform (LAMP) fuses three-dimensional (3D), real-time volumetric reconstructions of radiation sources with contextual information (e.g. LIDAR, camera, etc.) derived from the environment around the detector system. This information, particularly when obtained in real time, may be transformative for applications, includin… ▽ More Nuclear Scene Data Fusion (SDF), implemented in the Localization and Map** Platform (LAMP) fuses three-dimensional (3D), real-time volumetric reconstructions of radiation sources with contextual information (e.g. LIDAR, camera, etc.) derived from the environment around the detector system. This information, particularly when obtained in real time, may be transformative for applications, including directed search for lost or stolen sources, consequence management after the release of radioactive materials, or contamination avoidance in security-related or emergency response scenarios. 3D reconstructions enabled by SDF localize contamination or hotspots to specific areas or objects, providing higher resolution over larger areas than conventional 2D approaches, and enabling more efficient planning and response, particularly in complex 3D environments. In this work, we present the expansion of these gamma-ray map** concepts to neutron source localization. Here we integrate LAMP with a custom $Cs_2LiLa(Br,Cl)_6:Ce$ (CLLBC) scintillator detector sensitive to both gamma-rays and neutrons, which we dub Neutron Gamma LAMP (NG-LAMP). NG-LAMP enables simultaneous neutron and gamma-ray map** with high resolution gamma-ray spectroscopy. We demonstrate the ability to detect and localize surrogate Special Nuclear Materials (SNM) in real-time and in 3D based on neutron signatures alone, which is critical for the detection of heavily shielded SNM, when gamma-ray signatures are attenuated. In this work, we show for the first time the ability to localize, in 3D and realtime, a neutron source in the presence of a strong gamma-ray source, simultaneous and spectroscopic localization of three gamma-ray sources and a neutron source, and finally the localization of a surrogate SNM source based on neutron signatures alone, where gamma-ray data are consistent with background. △ Less

Submitted 16 August, 2019; originally announced August 2019.

3-D Radiation Map** in Real-Time with the Localization and Map** Platform LAMP from Unmanned Aerial Systems and Man-Portable Configurations

Authors: Ryan Pavlovsky, Andrew Haefner, Tenzing Joshi, Victor Negut, Keith McManus, Erika Suzuki, Ross Barnowski, Kai Vetter

Abstract: Real-time, meter-resolution gamma-ray map** is relevant in the detection and map** of radiological materials, and for applications ranging from nuclear decommissioning, waste management, and environmental remediation to homeland security, emergency response, and international safeguards. We present the Localization and Map** Platform (LAMP) as a modular, contextual and radiation detector sen… ▽ More Real-time, meter-resolution gamma-ray map** is relevant in the detection and map** of radiological materials, and for applications ranging from nuclear decommissioning, waste management, and environmental remediation to homeland security, emergency response, and international safeguards. We present the Localization and Map** Platform (LAMP) as a modular, contextual and radiation detector sensor suite, which performs gamma-ray map** in three dimensions (3-D) and in real time, onboard an unmanned aerial vehicle (UAV) or in a man-portable configuration. The deployment of an unmanned aerial system (UAS) for gamma-ray map** can be advantageous, as the UAS provides a means of measuring large areas efficiently and improving accessibility to some environments, such as multi-story structures. In addition, it is possible to increase measurement robustness through autonomous navigation, and to reduce radiation exposure to users as a result of the remote measurement. LAMP enables meter-resolution gamma-ray map** through Scene Data Fusion (SDF) [1], a capability that fuses radiation and scene data via voxelized 3-D Maximum Likelihood Expectation Maximization (MLEM) to produce 3-D maps of radioactive source distributions in real-time. Results are computed onboard LAMP while it is flying on the UAV and streamed from the system to the user, who can view the 3-D map on a tablet in real-time. We present results that demonstrate the SDF concept, including a set of UAS flights where a 133Ba source is localized at a test site in Berkeley, CA and a handheld measurement in Fukushima Prefecture, Japan where the distribution of radiocesium(137,134Cs) released from the accident of the Fukushima Daiichi Nuclear Power Plant is mapped. The reconstruction parameters used for each measurement were identical, indicating that the same algorithm can be used for both point or distributed sources. △ Less

Submitted 28 December, 2018; originally announced January 2019.

Teaching computational reproducibility for neuroimaging

Authors: K. Jarrod Millman, Matthew Brett, Ross Barnowski, Jean-Baptiste Poline

Abstract: We describe a project-based introduction to reproducible and collaborative neuroimaging analysis. Traditional teaching on neuroimaging usually consists of a series of lectures that emphasize the big picture rather than the foundations on which the techniques are based. The lectures are often paired with practical workshops in which students run imaging analyses using the graphical interface of spe… ▽ More We describe a project-based introduction to reproducible and collaborative neuroimaging analysis. Traditional teaching on neuroimaging usually consists of a series of lectures that emphasize the big picture rather than the foundations on which the techniques are based. The lectures are often paired with practical workshops in which students run imaging analyses using the graphical interface of specific neuroimaging software packages. Our experience suggests that this combination leaves the student with a superficial understanding of the underlying ideas, and an informal, inefficient, and inaccurate approach to analysis. To address these problems, we based our course around a substantial open-ended group project. This allowed us to teach: (a) computational tools to ensure computationally reproducible work, such as the Unix command line, structured code, version control, automated testing, and code review and (b) a clear understanding of the statistical techniques used for a basic analysis of a single run in an MRI scanner. The emphasis we put on the group project showed the importance of standard computational tools for accuracy, efficiency, and collaboration. The projects were broadly successful in engaging students in working reproducibly on real scientific questions. We propose that a course on this model should be the foundation for future programs in neuroimaging. We believe it will also serve as a model for teaching efficient and reproducible research in other fields of computational science. △ Less

Submitted 15 June, 2018; originally announced June 2018.