Abstract: Described is a knowledge-enhanced compressive imaging system. The system first initializes a compressive measurement basis set and a measurement matrix using task- and scene-specific prior knowledge. An image captured using the imaging mode of the dual-mode sensor is then sampled to extract context knowledge. The compressive measurement basis set and the measurement matrix are adapted using the extracted context knowledge and the prior knowledge. Task-relevant compressive measurements of the image are performed using the compressive measurement mode of the dual-mode sensor, and compressive reconstruction of the image is performed. Finally, a task and context optimized signal representation of the image is generated.

Abstract: Described is a method for image registration utilizing particle swarm optimization (PSO). In order to register two images, a set of image windows is first selected from a test image and transformed. A plurality of software agents is configured to operate as a cooperative swarm to optimize an objective function, and an objective function is then evaluated at the location of each agent. The objective function represents a measure of the difference or registration quality between at least one transformed image window and a reference image. The position vectors representing the current individual best solution found and the current global best solution found by all agents are then updated according to PSO dynamics. Finally, the current global best solution is compared with a maximum pixel value which signifies a match between an image window and the reference image. A system and a computer program product are also described.

Abstract: Described is a behavior recognition system for detecting the behavior of objects in a scene. The system comprises a semantic object stream module for receiving a video stream having at least two frames and detecting objects in the video stream. Also included is a group organization module for utilizing the detected objects from the video stream to detect a behavior of the detected objects. The group organization module further comprises an object group stream module for spatially organizing the detected objects to have relative spatial relationships. The group organization module also comprises a group action stream module for modeling a temporal structure of the detected objects. The temporal structure is an action of the detected objects between the two frames, whereby through detecting, organizing and modeling actions of objects, a user can detect the behavior of the objects.

Abstract: The present invention relates to a method for three-dimensional (3D) object recognition using region of interest geometric features. The method includes acts of receiving an implicit geometry representation regarding a three-dimensional (3D) object of interest. A region of interest (ROI) is centered on the implicit geometry representation such that there is at least one intersection area between the ROI and the implicit geometry representation. Object shape features are calculated that reflect a location of the ROI with respect to the implicit geometry representation. The object shape features are assembled into a feature vector. A classification confidence value is generated with respect to a particular object classification. Finally, the 3D object of interest is classified as a particular object upon the output of a statistical classifier reaching a predetermined threshold.

Abstract: Described is a system for visual object recognition using heterogeneous classifier cascades. Visual object recognition is one of the most critical tasks for video and image analysis applications. The present invention utilizes a cascade of classifiers, wherein each stage is dedicated to a certain task such as achieving high accuracy or reducing false alarms. The stages are then appropriately trained using either the training data or false alarm datasets, respectively. Additionally, the features that are employed by the classifier cascades are heterogeneous and complementary in that several types of features may be used. The system described herein has multiple applications in a variety of fields including automotive safety, factory automation, surveillance, force protection, and automatic target recognition.

Abstract: A multi-stage method of visual object detection is disclosed. The method was originally designed to detect humans in specific poses, but is applicable to generic detection of any object. A first stage comprises acts of searching for members of a predetermined general-class of objects (such as humans) in an image using a cognitive swarm, detecting members of the general-class of objects in the image, and selecting regions of the image containing detected members of the general-class of objects. A second stage comprises acts of searching for members of a predetermined specific-class of objects (such as humans in a certain pose) within the selected regions of the image using a cognitive swarm, detecting members of the specific-class of objects within the selected regions of the image, and outputting the locations of detected objects to an operator display and optionally to an automatic response system.

Abstract: Described is a system for object recognition in colorized point clouds. The system includes an implicit geometry engine that is configured to receive three-dimensional (3D) colorized cloud point data regarding a 3D object of interest and to convert the cloud point data into implicit representations. The engine also generates geometric features. A geometric grammar block is included to generate object cues and recognize geometric objects using geometric tokens and grammars based on object taxonomy. A visual attention cueing block is included to generate object cues based on 3D geometric properties. Finally, an object recognition block is included to perform a local search for objects using cues from the cueing block and the geometric grammar block and to classify the 3D object of interest as a particular object upon a classifier reaching a predetermined threshold.

Abstract: Described is a system for rapid directed area search utilizing particle swarm optimization. The system first extracts salient regions from an input image. The system then detects regions of interest from the salient regions utilizing particle swarm optimization, wherein a swarm of software agents, or particles, cooperate to locate an objective function optima, or region of interest, in an image. A set of local feature descriptors are then extracted from the image, wherein a local feature descriptor corresponds to a neighborhood surrounding a point of interest in a region of interest in the image. Additionally, the set of local feature descriptors are clustered hierarchically into a database so that a closest match between a new input image and a stored image can be determined. Finally, the matching regions of the two images are registered to align matching regions to allow detection of changes between the images.

Abstract: Disclosed is a method and system for generic object detection using block-based feature computation and, more specifically, a method and system for massively parallel computation of object features sets according to an optimized clock-cycle matrix. The method uses an array of correlators to calculate block sums for each section of the image to be analyzed. A greedy heuristic scheduling algorithm is executed to produce an optimized clock cycle matrix such that overlapping features which use the same block sum do not attempt to access the block at the same time, thereby avoiding race memory conditions. The processing system can employ any of a variety of hardwired Very Large Scale Integration (VLSI) chips such as Field Programmable Gate Arrays (FPGAs), Digital Signal Processors (DSPs) and Application Specific Integrated Circuits (ASICs).

Abstract: Described is a method for flexible feature adaptation and matching for object recognition in visual systems which incorporates evolutionary optimization. In the present invention, an analysis window is provided to select a portion of an input image to be analyzed for the presence or absence of an object. The analysis window is then divided into spatial regions, and a feature kernel function for each spatial region is selected and optimized. A feature value for each spatial region is calculated by finding a suitable location that generates the best matching features to a stored set using an optimization algorithm. The feature values are concatenated for the spatial regions to comprise a feature vector. Finally, the feature vector is processed by a classification algorithm, and a determination is made whether the object is present in the analysis window.

Abstract: Described is a system for multi-layered object detection which presents a unified way of processing an entire field-of-view (FOV) using cognitive swarms of software agents and classifier cascades by partitioning the FOV into layers and processing the closest layer first. A plurality of software agents operate as a cooperative swarm to search the first layer of the field-of-view to locate an objective function optima according to particle swarm optimization dynamics, wherein the objective function optima corresponds to a location of an object in the image in a layer of the field-of-view. The other layers are then sequentially swept to detect other objects in the FOV. In another aspect, the layers correspond to layers of increasing resolution in a hierarchical image pyramid. By using the cooperative swarm to search the coarser resolution layers first, objects can be detected more rapidly. A method and computer program product are also described.

Abstract: Disclosed is a method and system for generic object detection using block-based feature computation and, more specifically, a method and system for massively parallel computation of object features sets according to an optimized clock-cycle matrix. The method uses an array of correlators to calculate block sums for each section of the image to be analyzed. A greedy heuristic scheduling algorithm is executed to produce an optimized clock cycle matrix such that overlapping features which use the same block sum do not attempt to access the block at the same time, thereby avoiding race memory conditions. The processing system can employ any of a variety of hardwired Very Large Scale Integration (VLSI) chips such as Field Programmable Gate Arrays (FPGAs), Digital Signal Processors (DSPs) and Application Specific Integrated Circuits (ASICs).

Abstract: A safety monitoring system for a workspace area. The workspace area related to a region having automated moveable equipment. A plurality of vision-based imaging devices capturing time-synchronized image data of the workspace area. Each vision-based imaging device repeatedly capturing a time synchronized image of the workspace area from a respective viewpoint that is substantially different from the other respective vision-based imaging devices. A visual processing unit for analyzing the time-synchronized image data. The visual processing unit processes the captured image data for identifying a human from a non-human object within the workspace area. The visual processing unit further determining potential interactions between a human and the automated moveable equipment.

Abstract: A method and system for a directed area search using cognitive swarm vision and cognitive Bayesian reasoning is disclosed. The system comprises a domain knowledge database, a top-down reasoning module, and a bottom-up module. The domain knowledge database is configured to store Bayesian network models comprising visual features and observables associated with various sets of entities. The top-down module is configured to receive a search goal, generate a plan of action using Bayesian network models, and partition the plan into a set of tasks/observables to be located in the imagery. The bottom-up module is configured to select relevant feature/attention models for the observables, and search the visual imagery using a cognitive swarm for the at least one observable. The system further provides for operator feedback and updating of the domain knowledge database to perform better future searches.

Abstract: Described is a system for anomaly detection to detect an anomalous object in an image, such as a concealed object beneath a person's clothing. The system is configured to generate a subspace model for a normal class using training images. The normal class represents normal objects in a common class. The system receives a novel image having an object in the common class. A set of geometric landmarks are identified in the object in the novel image for use in registering the image. The novel image is registered by warping the image so that the geometric landmarks coincide in the novel image and the training images, resulting in a warped novel image having an object. Thereafter, the system determines if the object in the warped novel image is anomalous by measuring the distance of the warped novel image from the subspace model. Finally, if anomalous, an operator is notified accordingly.

Abstract: A safety monitoring system for a workspace area. The workspace area related to a region having automated moveable equipment. A plurality of vision-based imaging devices capturing time-synchronized image data of the workspace area. Each vision-based imaging device repeatedly capturing a time synchronized image of the workspace area from a respective viewpoint that is substantially different from the other respective vision-based imaging devices. A visual processing unit for analyzing the time-synchronized image data. The visual processing unit processes the captured image data for identifying a human from a non-human object within the workspace area. The visual processing unit further determining potential interactions between a human and the automated moveable equipment.

Abstract: A method, computer program product, and system for processing imagery is presented. The imagery is processed by receiving data regarding a scene (such as from a sensor monitoring a scene). The scene includes an object having a dimension. Flow vectors are computed from the data, while a flow histogram space is generated from the flow vectors. A line segment (with a length) is found within the flow histogram space. An object in the scene is associated with the length segment, and the dimensions of the object are estimated based on the length of the line segment.

Abstract: Anomaly prediction of battery parasitic load includes processing input data related to a state of charge for a battery and a durational factor utilizing a machine learning algorithm and generating a predicted start-up state of charge. Warnings are issued if the predicted start-up state of charge drops below a threshold level within an operational time.

Abstract: A vision-based system for automatically detecting the type of object within a specified area, such as the type of occupant within a vehicle is presented. The type of occupant can then be used to determine whether an airbag deployment system should be enabled or not. The system extracts different features, including wavelet features and/or a disparity map from images captured by image sensors. These features are then processed by classification algorithms to produce class confidences for various occupant types. The occupant class confidences are fused and processed to determine occupant type. In a preferred embodiment, image features from image edges, wavelet features, and disparity are used. Various classification algorithms may be implemented to classify the object. Use of the disparity map and/or wavelet features provides greater computational efficiency.

Abstract: An object recognition system is described that incorporates swarming classifiers. The swarming classifiers comprise a plurality of software agents configured to operate as a cooperative swarm to classify an object group in a domain. Each node N represents an object in the group having K object attributes. Each agent is assigned an initial velocity vector to explore a KN-dimensional solution space for solutions matching the agent's graph. Further, each agent is configured to search the solution space for an optimum solution. The agents keep track of their coordinates in the KN-dimensional solution space that are associated with an observed best solution (pbest) and a global best solution (gbest). The gbest is used to store the best solution among all agents which corresponds to a best graph among all agents. Each velocity vector thereafter changes towards pbest and gbest, allowing the cooperative swarm to classify of the object group.