Back to Company SearchSBIR-STTR Success: Kitware Inc.
The Air Force has a new way to share critical mission video footage that will bolster the confidence...
Company
Portfolio Data
KITWARE INC
UEI: DK6LPWMS5LP5
Number of Employees: 156
HUBZone Owned: No
Woman Owned: No
Socially and Economically Disadvantaged: No
SBIR/STTR Involvement
Year of first award: 1998
147
Phase I Awards
99
Phase II Awards
67.35%
Conversion Rate
$27,254,983
Phase I Dollars
$114,250,943
Phase II Dollars
$141,505,927
Total Awarded
Success Stories
See what our company has achieved through SBIR/STTR funding.
Awards
ARTIFICIAL INTELLIGENCE MULTIMODAL DAMAGE ASSESSMENT (AI-MDA)
Amount: $999,998 Topic: OSD252-001
Automating initial physical damage assessment (PDA) remains a challenge due to its wide variety of operating domains: sensor modalities, geographic regions, damage modalities, and imaging geometries. Labeling large amounts of training data to cover all of these conditions is not a feasible approach to creating automatic damage assessment tools. In particular, training data is usually completely unavailable for a given geographic domain before the onset of the damage event: few damaged buildings exist in a region before the event, even if there is forewarning of what the relevant region is. With Artificial Intelligence Multimodal Damage Assessment (AI-MDA), Kitware and Etegent will enable efficient creation of damage detectors for new operating domains with minimal human damage sample labeling. We utilize an architecture previously developed on AFRLÆs AI-BDA program and extend it to multiple modalities via a pre-trained remote sensing foundation model. This model can accept as input imagery either Capella SAR (our main focus) or EO data and outputs an information-rich embedding. Small damage detector heads can be trained on these embeddings to achieve effective performance with very few labeled training samples. In addition, we propose low-label and physics-informed methods to quantify model performance, assessment confidence, and operating bounds, which will aid further human review of predicted damage. SAR can be difficult to interpret for non-expert users, complicating the damage assessment process. To simplify results interpretation for users, we extend KitwareÆs RDWATCH software to visualize predicted damage with the help of EO imagery generated from input SAR images, and to provide input/output functionality in common data formats like GeoCOCO.
Tagged as:
SBIR
Phase II
2025
DOD
OSD
Hi-VALUE ATR
Amount: $1,999,999 Topic: A244-033
Automatic target recognition (ATR) remains a challenge due to the large variety of Army relevant sensors, sensing modalities, imaging conditions, and environments, nuanced differences in military targets, and the wide range of image scales over which a target may be imaged. Labeling vast amounts of training data to cover all of these conditions and targets is not a feasible approach. With Hi-VALUE ATR Kitware and Johns Hopkins University will enable an analyst to efficiently create new object detectors with zero or only a few object exemplars in the Linchpin Unclassified Environment by describing the object types in plain, descriptive text in addition to providing a few visual examples. We use a hybrid architecture, combining state-of-the-art vision language models (VLMs) with classical algorithmic components (e.g. Gaussian mixture models for motion detection, inpainting for synthetic occlusion generation, feature matching and homography estimation for image stabilization). Hi-VALUE ATR’s detectors will create results without lengthy retraining and quickly adapt to new domains (environments, targets, sensors, sensor modalities, weather conditions, etc.) because we build on modern VLMs and state-of-the-art few-shot learning methods. Preliminary results indicate that our approach is highly effective at ATR while requiring only limited training data.
Tagged as:
SBIR
Phase II
2025
DOD
ARMY
Complete Urban To Rural Balanced Streets By Artificial Intelligent Design (CURBS-AID)
Amount: $199,838 Topic: 24.2-CS1
Complete Streets planning is currently limited in scope by human bandwidth. Modeling and analyzing traffic infrastructure is time-consuming, and manually annotating transportation networks does not scale to large study areas. Kitware, along with NYU and MJ Engineering, proposes to build a Complete Streets AI assistant that can analyze infrastructure, answer queries, and make suggestions at scale. The foundation of this capability is an accurate traffic model; initially, we will build an accurate and complete digital traffic network capturing walkways and their interactions using state-of-the-art computer vision algorithms. In Phase I, we will extend our prior work to automatically extract walkway networks from overhead image surveys, incorporating computer vision techniques on ground-level imagery to enhance accuracy and provide finer details such as curbs and sidewalk conditions. If successful, our approach will automate the generation of walkway networks at a national scale with minimal cost. Phase II will develop an LLM assistant to interact with planners through a retrieval augmented generation framework that can respond to queries by indexing the traffic network and generating proposals. Finally, we will use our team’s web-based visualization tools to create an interface for planners to interact with the assistant and visualize proposed network alterations.
Tagged as:
SBIR
Phase I
2025
DOT
FIELDViewR: Field-Focused Inspection, Engineering, and Logistics Digital Viewer for Roadways
Amount: $206,500 Topic: DOT-25-FH1
Large-scale infrastructure projects rely on open standards such as Industry Foundation Classes (IFC) to streamline collaboration between architects, engineers, contractors, and owners. However, efficiently visualizing, managing, and monitoring complex IFC-based models in real-time remains challenging, especially when incorporating field data such as LiDAR scans and ground-based photos. Kitware, Inc. proposes FIELDViewR, a web-based BIM/IFC viewer that supports streaming data, geospatial alignment, and on-site progress tracking with reality-enhanced insights. Phase I prototype development will focus on core capabilities for interactive visualization of IFC models over the web, employing level-of-detail strategies for smooth streaming regardless of dataset size. The project will integrate basic geospatial alignment and camera calibration, enabling seamless overlays between digital infrastructure models and real-world coordinates. This foundation will support preliminary reality capture features that let field personnel compare site captures against design models for progress tracking. The platform will establish a foundation for digital-physical comparison in Phase II, enabling efficient real-time comparison of physical structures with digital counterparts. Ultimately, FIELDViewR promises to reduce rework, accelerate decision-making, and improve site safety by unifying BIM data, geospatial context, and real-world observations in one easy-to-use environment.
Tagged as:
SBIR
Phase I
2025
DOT
Composable Digital Twins for Science Network Infrastructures using Parallel Discrete Event Simulation
Amount: $1,149,997 Topic: C57-03a
Wide area network infrastructures, such as the Department of Energy’s (DOE) ESnet, are critical for moving large amounts of data between experimental facilities and data centers, presenting significant challenges related to resource allocation, provisioning, and performance optimization. Testbeds, like the ESnet testbed, are commonly used to evaluate and optimize network performance, but physical testbeds have inherent limitations in scalability, cost, and flexibility. A network digital twin, a virtual replica of a physical network, provides a powerful tool to overcome these limitations. It enables experimentation at scale, exploration of ”what-if” scenarios, and long-term simulations that would be difficult or impossible with physical testbeds alone. Network simulators have been widely used in network research and are well-suited for integration into a network digital twin. However, these tools are typically slow in rapid research and development environments, as even simulating a millisecond of network traffic can take hours when maintaining high fidelity. To address this, we propose developing a digital twin framework that guides users through the entire network simulation workflow. The tool will integrate several network simulators, each focusing on different aspects of network behavior, such as hardware and network protocols, to provide a comprehensive understanding of complex network infrastructures. To accelerate simulation, we will employ a hybrid modeling approach, combining high-fidelity simulation with machine learning models that can “fast-forward” through parts of the simulation. In Phase I, a prototype of the digital twin framework and an interactive visualization dashboard was developed to help users understand simulation results. Additionally, an initial machine learning model was created to predict switch behavior of the ESnet testbed. In Phase II, the framework will be expanded to include a user-friendly application that guides users through the process of configuring models, network topologies, simulation ensembles, and visualizing results. Machine learning models will also be enhanced to further speed up simulations, with the integration of foundation models and federated learning techniques to preserve data privacy. After Phase II, the framework will be expanded beyond research network providers, such as the DOE, to support cloud service providers, which manage applications requiring high throughput and low latency. These applications, such as streaming analytics services, are used for real-time decision making, where increased latency can result in financial consequences for users of cloud services. The proposed framework has the potential to improve the design and operation of these systems by providing insights into network performance optimization, planning, and resource management.
Tagged as:
SBIR
Phase II
2025
DOE
MetaMosaic: Metadata Guided Geospatial Mosaics from FMV
Amount: $999,998 Topic: OSD242-D001
The construction of geo-registered mosaic images from full motion video (FMV) can provide significant value to imagery analysts. When FMV provides high resolution video it also lacks in spatial context. A mosaic of video frames can help provide that context while maintaining high resolution. However, producing an accurate mosaic from operational FMV can be extremely challenging. Operators frequently pan or zoom the sensor rapidly and change between sensing modalities. FMV also frequently has metadata overlays burned into the pixels on screen, moving foreground objects, compression, and transmission artifacts. Kitware proposes MetaMosaic to improve mosaic generation by leveraging the metadata in the stream, either key-length-value (KLV) encoded or recognized from onscreen text. We not only read metadata values from the screen but also remove those metadata overlays so they do not produce artifacts in the mosaic. Our approach combines sub-pixel accurate image feature tracking with global location constraints from metadata to optimize a high quality mosaic even under challenging conditions. We propose a baseline 2D image registration as well as a full 3D solution for more complex scenes and camera motion. To remove barriers, we propose our FMV mosaic generation tool and all source code with unlimited rights to the government.
Tagged as:
SBIR
Phase II
2025
DOD
OSD
StreamLine: An AI-Enabled Open Source Publication Platform to Revolutionize Scientific Communication
Amount: $140,000 Topic: N242-091
The future of scientific communication is moving towards openness, fostering highly collaborative environments and extending beyond traditional publication artifacts with the rise of data-driven AI. This shift necessitates platforms that support open science and offer the customizations needed for current and future trends in scientific research. However, current systems are often rigid, tailored to specific workflows, and difficult to customize, falling short in managing diverse forms of scientific communication, such as datasets, algorithms, and other artifacts. Moreover, these systems fail to leverage recent advancements in Artificial Intelligence (AI), particularly the large language models (LLMs) that have driven the success of tools like ChatGPT. To address these gaps, Kitware and the University of Illinois Urbana-Champaign propose developing StreamLine, a novel AI-enabled publication platform with an open software infrastructure supporting modern scientific and technical communication. StreamLine will leverage LLMs and augment them with domain knowledge via Retrieval-Augmented Generation (RAG) to automate the inner workings of the publication process while creating a highly engaging environment for authors, reviewers, and other stakeholders through web interfaces. It uses Git for versioning of the contents and Postgres for indexing the metadata. A key feature of the platform will be the ability to support customizable publication workflows around a diverse range of content and artifact types. Towards this goal, in Phase I, we will 1) conduct a comprehensive survey of existing software and workflows adopted by current publication systems for geosciences, 2) build the Phase I StreamLine prototype using open-source standards and tools to drive the publication process for journal articles, 3) extend and integrate pre-trained open-source LLMs and augment them with domain knowledge to find and suggest related work, and 4) gather community and stakeholder feedback. In Phase II, we plan to iterate and enhance the software by expanding its capabilities to support a broader range of technical and scientific communication use cases and implementing complete end-to-end testing. We will collaborate closely with Naval Research, universities, and other scientific publication partners to further refine and improve the software. Our developed platform and models will be open source and freely available, subject to government permission, actively encouraging community review and contributions.
Tagged as:
SBIR
Phase I
2025
DOD
NAVY
Bridging the Gap in Prolonged Field Care: Integrated Visualization and Predictive Modeling for Extended Combat Casualty Training
Amount: $249,969 Topic: A254-029
As the domain of conflict has shifted the need to prioritize not only the “golden hour” of Tactical Combat Casualty Care (TCCC), but also the extended timeline associated with Prolonged Casualty Care (PCC) has become clearer. Injury and wound management over hours to days presents unique challenges not covered in the TCCC training. Current training modalities, including VR, AR, XR, and moulage, have limitations for representing the time progression of wounds. The combined skills of the Kitware and Exonicus team are well positioned to address these challenges with experience developing predictive medical models, medical training content, and medical visualization. Our solution proposes combining state-of-the-art technology and algorithms, including flexible screen technology, the Pulse Physiology Engine for dynamic patient feedback over time, tablet control of the medical scenario and wound visualization, and computer vision algorithms to scale and visualize the wound and underlying anatomy on live training participants and manikins. In Phase I, we will develop a prototype of the primary technological components and features for feasibility testing. A feasibility assessment will be provided to identify gaps, propose future directions and the refinement of the various components and features, and obtain stakeholder feedback.
Tagged as:
SBIR
Phase I
2025
DOD
ARMY
Real-Time Automated Patient Identification and Documentation - TC3 (RAPID-TC3)
Amount: $1,300,000 Topic: DHA254-DP002
The limited medics available on the battlefield are tasked not only with treating all patients but also with documenting each patient's injuries and care. This proposal presents RAPID-TC3 to assist medics by using artificial intelligence (AI) to automate tactical combat casualty care (TC3) reporting. RAPID-TC3 consists of a headset-mounted camera that is robust to challenging battlefield conditions, including bright/low-lighting and diverse weather conditions. RAPID-TC3 uses COTS stereo cameras to process the scene and generate its 3D representation. We detect all casualties, medical equipment, and medical procedures performed by the medic on each patient. Using depth from stereo, we create 3D skeletons, body part segments, and body shapes for each casualty. We track and re-identify each casualty as the patients and medics move throughout the scene. Per-module updates are done with varying frequency, to balance accuracy and computational efficiency. By mapping the detected medical devices and interventions to each patient's body, we are able to generate TC3 report data for each casualty. Our system can run for 2 hours, and all computing is done on the edge.
Tagged as:
SBIR
Phase II
2025
DOD
DHA
GU3SS: Generative Unbiased 3D Semantic Segmentation
Amount: $999,998 Topic: OSD233-001
3D models are commonly generated from both multiview satellite and FMV sources using photogrammetric methods. However, such models lack semantic labels (e.g. segmentation of buildings, roads, vegetation, vehicles, etc.) needed for further analytics. Prior work relies largely on discriminative models to classify 3D surface points using local context, but they also have difficulty learning complex relationships between objects that generalize to new domains. Recent work in Vision-Language Foundation Models (VLMs) trained on a combination of imagery and language has shown great power in modeling and generating large-scale imagery consistently and meaningfully. It has also been shown that generative models used in semantic segmentation can generalize better to new domains than discriminative models. Kitware proposes to leverage these recent findings to provide semantic segmentation of 3D surfaces via the fusion of multiple generated 2D segmentation maps. We will leverage the fact that these 3D models are derived from 2D imagery to adapt and fine-tune existing VLMs to generate semantically meaningful segmentation in 2D before fusing into consistent 3D surface labels. The result will be 3D segmentation and detection results that generalize far better to new environments.
Tagged as:
SBIR
Phase II
2025
DOD
OSD