Agency: Department of Defense | Branch: Army | Program: SBIR | Phase: Phase II | Award Amount: 773.82K | Year: 2011
Fully automated approaches to target detection are efficient at processing large amounts of data but often rely heavily on the training data and the model employed. Training data and modeling assumptions can be violated in the operational environments where the algorithms are applied. Alternatively, human manual target detection is accurate and adaptable due to the human ability to interpret data within its wider context. However, operational constraints and the overwhelming amount of data from modern sensors frequently preclude a fully manual approach to target detection. The SIG human-in-the-loop (HIL) active learning (AL) framework allows the operators to guide the training of the automated detection algorithm when the training and testing data statistics are mismatched. In this structured framework, the algorithm cues the operator using two specific criteria: detections that are high probability targets and detections that are highly informative for improving the classifier performance. The operator provides labels for the cued detections, and the new label information is used to retrain the classifier and improve the performance of the algorithm. At the conclusion of this Phase II effort, SIG will deliver a C/C++ implementation of the HIL/AL architecture that is ready for integration and test on the Shadow IED program.
Agency: Department of Defense | Branch: Air Force | Program: SBIR | Phase: Phase II | Award Amount: 745.07K | Year: 2010
A probabilistic ATR framework is proposed to exploit coincident multi-aspect radar and EO video data for target detection, tracking, and classification/identification. The mathematical framework is constituted by four principal components with a particular focus on exploitation of 3D information from radar scattering: 1) extraction of features indicative of shape and structure from radar waveforms, 2) tracking of targets from coincident EO and radar data to estimate target-sensor pose, 3) estimation of 3D target representations from radar data, and 4) development of pattern recognition algorithms to provide target classification and identification. This framework will exploit all sources of information in the EO/radar data. The probabilistic framework supports rigorous characterization of the uncertainty associated with each component as well as propagation to subsequent dependent components. This uncertainty is ultimately captured in the target classification and identification results. The products from the proposed research offer the potential for significant further improvements in target ID performance for realistic operational environments. SIG will collaborate with AFRL to identify the appropriate measured data sets to demonstrate and evaluate the framework. Potential data sets include the AFRL Layered Sensing collect (Angel Fire EO with GOTCHA SAR) and/or Bluegrass data (Constant Hawk EO with JSTARS GMTI). BENEFIT: Layered sensing and multi-sensor fusion, medical imaging, traffic analysis, multi-sensor security, next-generation ISR
Agency: Department of Defense | Branch: Navy | Program: SBIR | Phase: Phase II | Award Amount: 734.27K | Year: 2010
In the proposed Phase II program, the methods developed and implemented during Phase I research will be fully integrated within a common Bayesian in situ learning framework. We have developed several Bayesian classifiers, to which we will apply label acquisition and label confidence techniques. Additionally, we will extend the in situ learning framework to include multi-task learning. Previously collected sensing data are often available from different sensors or environments. Not all data are related, however the potential exists to share information between related tasks and exploit the contextual information of previous tasks. The current in situ learning process is inherently myopic; the algorithm identifies the single most-informative data sample. The ability to select multiple samples without relearning the classifier can increase computational efficiency and maximize analyst workload. Based on the theory of submodular functions, non-myopic in situ learning techniques for subset selection will be developed and integrated into the Bayesian framework. Finally, new statistical embedding technology will be investigated that allows an analyst to synthesize data for training and to augment the label acquisition process. A low-dimensional embedded space may be visualized, and any location on the manifold can be recreated in the original high-dimensional space.
Agency: Department of Defense | Branch: Navy | Program: SBIR | Phase: Phase I | Award Amount: 149.90K | Year: 2013
Signal Innovations Group proposes a hierarchical Bayesian approach for non-linear dimensionality reduction that addresses three key challenges: learning a reversible mapping from a high-dimensional observed space to a low-dimensional embedded space, learning the dimension of the embedded space, and generating new high-dimensional data for a given location in the embedded space. The proposed generative approach is statistical and jointly learns the probabilistic reversible mapping and the dimension of the embedded space. The proposed approach also enables new high-dimensional data to be embedded in a previously learned low-dimensional space. A hierarchical Bayesian method is also proposed to learn a non-linear dynamic model in the low-dimensional space, allowing joint analysis of multiple types of dynamic data, synthesis of new dynamic data in the low-dimensional space, and mapping synthesized data to the high-dimensional observation space. The models are designed to uncover the relevant characteristics and structure of data through non-linear dimensionality reduction, which enables a human analyst to identify and explore the characteristics in the low-dimensional manifold space and generate new unobserved high-dimensional data.
Agency: Department of Defense | Branch: Air Force | Program: SBIR | Phase: Phase I | Award Amount: 149.94K | Year: 2012
ABSTRACT: The Air Force has invested considerable resources into collecting and synthesizing radar data for training and testing ATC/R systems. Significant cost savings may be realized if these existing datasets may be leveraged for training new sensors and modalities. Signal Innovations Group offers a new paradigm for automatically identifying statistically salient features for ATR systems from combined sources of existing surrogate and limited operational data. This approach departs from conventional techniques that attempt to compensate for numerous sources of degradation, through pre-processing, noise estimation, modeling and manual intervention in order to obtain perfect HRR template matches. Saliency analysis identifies sparse subsets of features (e.g. HRR range bins) that are both statistically significant for ATR and robustly manifested in data. Salient features have been shown to provide superior classification performance compared to full-dimensional HRRs. Additionally, SIG proposes migrating away from conventional HRRs to simple, compact sets of physics-based features, derived from EM phenomenology, which may be extracted independently from both SAR and MTI. This paradigm avoids reliance on complex pre-processing to compensate for distortion by utilizing statistical inference techniques to identify robust phenomenology. Adverse phenomena (e.g. multi-bounce or shadowing) are highly variable and will be rejected by the saliency analysis. BENEFIT: The successful program will result in a capability to automatically leverage data across multiple sensors and modalities for ATR development. A common database of physics-based features will be developed across families of sensors. This will reduce the costs of training new ATR systems and decrease the time required to deploy new capabilities. The Bayesian framework naturally supports the fusion of information from alternative sensing modalities such as optical, infrared, or hyper-spectral. The physics-based features in the radar regime may be combined with corresponding physics-based features in these alternate regimes, through higher-level Bayesian processes, to improve overall ATR performance. The sequential Bayesian inference framework for identifying salient features has potential extensions in both military and commercial applications. This framework may be applied to train new ATR systems, with very limited characterized data, using surrogate datasets from existing sensors. This applies to new sensors, including EO and IR, developed for military or geospatial applications as well as for medical imaging and diagnostic applications. For example, this framework may be applied towards improved tissue characterization and disease detection with medical imaging systems, automated facial recognition systems, genetic/protein structure and function determination for bioinformatics analysis, and next-generation internet search engine development.
Agency: Department of Defense | Branch: Navy | Program: SBIR | Phase: Phase II | Award Amount: 499.99K | Year: 2014
Signal Innovations Group (SIG) proposes a Phase II program that addresses three primary capabilities for applications involving high-dimensional observed data: reversible nonlinear data dimensionality reduction, static and dynamic synthesis of data in the observed space based on a low-dimensional latent space, and intuitive and meaningful human interaction with the data in both the observed and latent spaces. Each of these three capabilities will be addressed in the Phase II Base effort through the development of models and algorithms, visualization techniques, and a prototype user interface. Additionally, proposed Phase II Options 1 and 2 will focus on maturing and optimizing the technology and software, ultimately leading to transition and integration with a targeted program of record.
Agency: Department of Defense | Branch: Air Force | Program: SBIR | Phase: Phase II | Award Amount: 749.95K | Year: 2014
ABSTRACT: The Phase II program will develop and mature a new salient physics-driven solution for CID feature design and classification to support onboard CID and decision fusion for remotely piloted vehicles. Saliency analysis will be used to develop databases of compact, simple geometric features derived from known, discriminative, and robust scattering physics. A new Bayesian probabilistic classifier be developed and validated for SAR and HRR modes that exploit a hybrid combination of conventional sparse, data-driven signature features and efficient physics-based features. Uncertainty will be propagated from feature extraction to decision. Target libraries will include both civilian and military vehicles. The classifier will be developed in a hierarchical fashion that utilizes as little target information as possible in order to achieve CID objectives. Primarily, efficient geometric features will be used for most CID tasks. When increased CID fidelity or reliability are required (e.g., high value targets, or known confusors), sparse-signature salient feature exploitation techniques will be used to supplement the classifier with additional information. The salient physics-driven solution will address the limited processing and memory challenges associated with onboard CID accuracy and confidence. Furthermore, salient physics-based feature design is an enabling technology for non-cooperative target modeling. BENEFIT: The Phase II program is designed to address the efficiency and sustainability issues associated with the development, operation, and maintenance of current non-cooperative ATR technology. The products of the proposed program will ultimately lead to a low-cost, quick turn-around solution for target insertion into CID databases, at a significant savings compared to conventional signature database enablers. The selection of salient, physics-based features will reduce the template/database dimensionality for multi-phenomenology CID. The databases of compact feature sets identified by saliency analysis will provide CID accuracy and reliability. The proposed program represents a significant change in CID design practices; tasking under the Phase I resulted in a proof of concept that addresses the system requirements of and offers risk reduction to anticipated future AFRL requirements. In Phase II, SIG will further develop the technical maturity of this architecture, explore algorithm migration paths and requirements to operational use, and demonstrate performance in a lab environment.
Agency: Department of Defense | Branch: Army | Program: SBIR | Phase: Phase II | Award Amount: 599.96K | Year: 2013
The detection and tracking of dismounts and the identification of threatening dismounts are critical challenges for ensuring mission success and troop safety when conducting tactical vehicle operations in hostile environments. Vehicle operators must focus on multiple tasks simultaneously: navigation, collision avoidance, location of possible IED threats, and identification of dismount threats. Fatigue and distractions affect performance when mission lengths extend over long periods of time. Automated sensor systems offer the opportunity to perform dismount threat monitoring to reduce the surveillance burden of vehicle operators and to maintain an acceptable risk level. In this Phase II effort, Signal Innovations Group (SIG) proposes to build a software package that leverages SIGs existing tracking technology and builds on the successes of the Phase I program to provide automated, geo-referenced detection and tracking of dismounts as well as the identification of threatening dismounts from vehicle-mounted EO/IR full motion video (FMV) sensor systems. Furthermore, SIG proposes to output this geo-referenced track and threat metadata into a database that may be queried by other visualization and processing tools, transforming tactical EO/IR FMV sensors into non-traditional intelligence, surveillance, and reconnaissance (ISR) assets for improving battlefield situational awareness and assisting in the identification and remediation of asymmetric threats.
Agency: Department of Defense | Branch: Air Force | Program: SBIR | Phase: Phase I | Award Amount: 149.81K | Year: 2013
ABSTRACT: Current automatic target recognition (ATR) training processes require expensive data collections or extensive, high fidelity target modeling and validation whose costs and lead times will limit the ongoing sustainability of ATR target databases. Radar based systems for combat identification (CID) suffer from sustainability issues due to the extreme complexity of the target databases and the high costs and latency of incorporating new targets to meet evolving mission needs. In order to enable sustainable, reliable radar CID through salient physical features, SIG proposes to leverage existing HRR-based saliency technology to develop a knowledge base of target class and aspect dependent geometric features (such as the distance between critical scattering structures such as a bumper and windshield) from existing data for compact, robust CID. This simple and robust physical feature domain will be used to train a novel probabilistic classifier architecture that characterizes the uncertainty of target decisions. SIG will emphasize the selection of physics-based features that are relevant across a wide range of sensing modalities (HRR, SAR, EO), expanding the availability of target training data and facilitating future development and capabilities. These salient features enable sustainable development, operation, and maintenance of a compact, robust, and discriminative CID database. BENEFIT: A successful Phase I will result in a CID ATR framework that addresses the efficiency and sustainability issues associated with the development, operation and maintenance of current non-cooperative ATR technology. The proposed method provides a low-cost, quick turn-around solution for target insertion into ATR databases, at a significant savings compared to conventional signature database enablers. The selection of salient, physics-based features will reduce the template/database dimensionality for multi-phenomenology ATR by replacing image/signature template databases with compact feature sets. The proposed Phase I results in a proof of concept that addresses the system requirements of and offers risk reduction to future AFRL efforts.
Agency: Department of Defense | Branch: Air Force | Program: SBIR | Phase: Phase II | Award Amount: 1.05M | Year: 2013
ABSTRACT: This program will develop salient feature analysis and saliency cueing capabilities for SAR-based exploitation and demonstrate this technology with operationally relevant data and classifiers. This saliency technology will immediately enable the development of exploitable, robust, compact feature databases for sustainable radar CID. This new SAR-based saliency capability will complement existing HRR-based saliency technology by providing either independent or joint discovery of sparse sets of robust, exploitable features across these two common radar sensing modalities, in support of air-to-air or air-to-ground moving and stationary CID missions. This framework provides a balance between a compact target representation and target generalization across in-class variations. The underlying probabilistic model captures the uncertainty inherent to the training process due to limited or noisy data, and propagates these probabilities in a mathematically rigorous manner to downstream processes. The proposed work for this program will focus on near-term capabilities of sustainable radar CID database development and will lay the fundamental groundwork for potential fusion of radar and EO features based upon saliency and target geometry. BENEFIT: Phase II will result in a combined salient feature analysis, saliency cueing, and classification performance analysis capability for both HRR and SAR radar modalities integrated into POSSIBLE. This capability will enable the automatic identification of compact exploitable radar signature features over fewer target aspect states for more efficient and sustainable CID databases. While the salient feature analysis framework will be leveraged on the proposed program for classification with radar sensor modalities, the framework easily generalizes to other data modalities and exploitation applications, such as detection, recognition, identification, and general data categorization. Medical device manufacturers are developing cutting-edge sensors and equipment that is leading to revolutionary advances in non-invasive diagnosis of a variety of diseases, especially cancer. These devices have the ability to gather a significant amount of data, much more than a physician or technician can handle alone. Additionally, commercial satellite imagery providers are developing sensors that collect extremely high-resolution imagery. While more pixels generally provide more information, large amounts of data also lead to increased burden on image analysis, storage, and query. Therefore, automated processing, including rigorous understanding of salient features, is necessary for realizing the full potential of these new medical devices and satellite imagery. SIG is actively engaged with sensor and data providers in each of these industries and will leverage these relationships to commercialize the products of this Phase II SBIR.