Abstract: A device (e.g., an image sensor, camera, etc.) may identify a camera lens and color filter array (CFA) sensor used to capture an image, and may determine filter parameters (e.g., a convolutional operator) based on the identified camera lens and CFA sensor. For example, a set of kernels (e.g., including a set of horizontal filters and a set of vertical filters) may be determined based on properties of a given lens and/or q-channel CFA sensor. Each kernel or filter may correspond to a row of a convolutional operator (e.g., of a restoration bit matrix) used by an image signal processor (ISP) of the device for non-linear filtering of the captured image. The corresponding outputs from the horizontal and vertical filters (e.g., two outputs of the horizontal and vertical filters corresponding to an input channel associated with the CFA sensor) may then be combined using a non-linear classification operation.

Abstract: A device (e.g., an image sensor, camera, etc.) may identify a camera lens and color filter array (CFA) sensor used to capture an image, and may determine filter parameters (e.g., a convolutional operator) based on the identified camera lens and CFA sensor. For example, a set of kernels (e.g., including a set of horizontal filters and a set of vertical filters) may be determined based on properties of a given lens and/or q-channel CFA sensor. Each kernel or filter may correspond to a row of a convolutional operator (e.g., of a restoration bit matrix) used by an image signal processor (ISP) of the device for non-linear filtering of the captured image. The corresponding outputs from the horizontal and vertical filters (e.g., two outputs of the horizontal and vertical filters corresponding to an input channel associated with the CFA sensor) may then be combined using a non-linear classification operation.

Abstract: Systems and methods are disclosed for processing spectral imaging (SI) data. A training operation estimates reconstruction matrices based on a spectral mosaic of an SI sensor, generates directionally interpolated maximum a-priori (MAP) estimations of image data based on the estimated reconstruction matrices. The training operation may determine filter coefficients for each of a number of cross-band interpolation filters based at least in part on the MAP estimations, and may determine edge classification factors based at least in part on the determined filter coefficients. The training operation may configure a cross-band interpolation circuit based at least in part on the determined filter coefficients and the determined edge classification factors. The configured cross-band interpolation circuit captures mosaic data using the SI sensor, and recovers full-resolution spectral data from the captured mosaic data.

Abstract: Systems and methods are disclosed for processing spectral imaging (SI) data. A training operation estimates reconstruction matrices based on a spectral mosaic of an SI sensor, generates directionally interpolated maximum a-priori (MAP) estimations of image data based on the estimated reconstruction matrices. The training operation may determine filter coefficients for each of a number of cross-band interpolation filters based at least in part on the MAP estimations, and may determine edge classification factors based at least in part on the determined filter coefficients. The training operation may configure a cross-band interpolation circuit based at least in part on the determined filter coefficients and the determined edge classification factors. The configured cross-band interpolation circuit captures mosaic data using the SI sensor, and recovers full-resolution spectral data from the captured mosaic data.

Abstract: A method and system for collaborative communications is described. In one embodiment, a central virtual reality communications environment is created. A plurality of client communication devices are connected to the central virtual reality communications environment. Each one of the connected plurality of client communication devices are represented as an avatar present in the central virtual reality communications environment. An uploaded data object is received from any one of the connected plurality of client communication devices. Finally, the data object is displayed in the central virtual reality communications environment to the connected plurality of client communication devices.

Abstract: A signal processing method and power amplifier device are disclosed. The method may include receiving a signal to be transmitted, decomposing an original signal into a plurality of smaller constant-amplitude signals, wherein a vector sum each of the smaller constant-amplitude signals equals the original signal, amplifying the smaller constant-amplitude signals by an amplification factor using a plurality of amplifiers, wherein one or more of the plurality of amplifiers are enabled based on the amplitude of the original signal, combining the amplified smaller constant-amplitude signals into the original signal, the original signal being amplified by the amplification factor, wherein the amplified original signal is transmitted.

Abstract: A recursive lambda rule engine (114, 302) includes a first multiplier (204) that sequentially multiplies each of series of inputs by a nonlinearity determining parameter and supplies results to a second multiplier (214) that multiplies the output of the first multiplier (204) by a previous output of the engine (114, 302). A three input adder (220, 228) sequentially sums the output of the second multiplier (214), inputs from the series of inputs, and the previous output of the engine (114, 302). A shift register (244) is used to feedback the output of the engine (114, 302) to the three input adder (220, 228) and second multiplier (214). A MUX (234) is used to route an initial value through the shift register (244) for the first cycle of operation.

Abstract: A risk management system (100) for first responders (102) equips first responders with physiological and environmental sensors (104) in order to track the stress conditions and stress responses of the first responders. Data from the sensors is digitized and filtered and processed through a feature extraction system (202, 704) that may include noise filters, (710) derivative filters (706), integrators (708) and derived signal generators (712) that use stored data on first responders, before being fed into a pattern recognition system (402, 900) that assesses the risk state of the first responder. Data from multiple first responders in a team can be aggregated to obtain team risk assessments.

Abstract: A method (1500) and apparatus (700, 2300, 2400) for aggregating two or more input signals with a versatile reconfigurable signal aggregator. The aggregator (700, 2300, 2400) is reconfigured by adjusting a control signal ?, and can emulate a range of union type signal aggregators, a range of intersection type signal aggregators, and a continuum of functions between the two, including a signal averager. The versatility of the aggregator (700, 2300, 2400) allows systems in which the aggregator (700, 2300, 2400) is incorporated to be highly adaptable, and thereby fosters improved machine learning.

Abstract: Disclosed is a communications device that monitors the status of its communications links. When more than one communications link is available, an “appropriate” link is selected for use by an application running on the device. The links are continually monitored, and, as circumstances change, the link selection can also change. In some embodiments, a link is “appropriate” when it is available to transfer data and when it satisfies certain “hard” rules set by the application or by a user of the communications device. Some embodiments monitor the behaviour of the communications device and of its user and capture that behaviour in “soft” rules that are applied when selecting an appropriate link and when the hard rules leave the choice open. When multiple links transmit on the same frequency, they may be supported “virtually simultaneously” by time-slicing the actual data transmissions. Unneeded links may be shut down to save battery power.

Abstract: Signal processing networks (700, 800, 1008, 1010, 1012) that include a configurable infinite logic aggregator (100) that can be configured as an infinite logic AND gate and infinite logic OR gate or as other gates along a continuum of function between the two by adjusting control signal magnitudes and a configurable infinite logic signal inverter (500) are provided. A method of designing such networks that includes a genetic programming program (1802) e.g., a gene expression programming program (1600), for designing the network topology, in combination with a numerical optimization (1804), e.g., a hybrid genetic algorithm/differential evolution numerical optimization (1700) for setting control signal values of the network and optionally other numerical parameters is provided.

Abstract: Computer programs (600, 700, 800, 900, 1000) and a programmed computer (1100) for automatically generating computer programs (i.e. sequences of instructions) are provided. The computer programs (600, 700, 800, 900, 1000) use Hidden Markov Models (400, 500) to generate sequences of program tokens, e.g., Gene Expression Programming chromosomes (100). Parameters of the Hidden Markov Models (400, 500) are numerically optimized, for example, by Differential Evolution with a goal of increasing the fitness of automatically generated programs.

Abstract: An nonlinear digital signal processing filter (100, 200, 1100, 1308, 1310, 1312, 1346, 1604) maintains a magnitude ordering for successive windows of signal samples. A set of filter density generator values [f1, f2, f3 . . . fj . . . fndensities] are used according to the ordering in a recursion relation that computes successive values of a set function over the set of filter density generator values. The recursion relation involves an adjustable nonlinearity defining parameter ?. The values are normalized by dividing by a largest of the values, and differences between successive values are taken. An inner product between each window of signal values (used in order according to magnitude) and the adaptive differences is a filtered signal sample.

Abstract: A Q-Filter is a reconfigurable technique that performs a continuum of linear and nonlinear filtering operations. It is modeled by unique mathematical structure, utilizing a function called the Q-Measure, defined using a set of adjustable kernel parameters to enable efficient hardware and software implementations of a variety of useful, new and conventional, filtering operations. The Q-Measure is based on an extension of the well-known Sugeno ?-Measure. In order to optimize the Q-Filter kernel parameters, the value of an error function is minimized. The error function is based on difference between the filtered signal and target signal, with the target signal being a desired result of filtering.

Abstract: Computer programs (600, 700, 800, 900, 1000) and a programmed computer (1100) for automatically generating computer programs (i.e. sequences of instructions) are provided. The computer programs (600, 700, 800, 900, 1000) use Hidden Markov Models (400, 500) to generate sequences of program tokens, e.g., Gene Expression Programming chromosomes (100). Parameters of the Hidden Markov Models (400, 500) are numerically optimized, for example, by Differential Evolution with a goal of increasing the fitness of automatically generated programs.

Abstract: A multi-sensor system includes a first analog sensor sub-system, a second analog sensor sub-system, and a system for synchronizing the outputs of the first and second analog sub-systems. Each analog sensor sub-system includes a sensor that produces an analog output. Each sensor is coupled to analog circuitry that processes the output from the sensor. The system for synchronizing the outputs of the first and second analog sensor sub-systems simultaneously inserts a marker into the outputs of the first and second analog sensors. Then, the outputs of the analog circuitry of the first and second analog sub-systems are synchronized based upon the marker. The marker signal may be produced using a Barker sequence signal generator.

Abstract: A first negotiation (142) is performed between a first agent (112) and a second agent (122). Responsive to the outcome of the first negotiation (142), a first asset is selectively purchased from the second agent (122). A second negotiation (148) is performed with a third agent (136), and, responsive to the outcome of the second negotiation (148), a second asset controlled by the first agent (112) is selectively sold to the third agent (136).

Abstract: A Support Vector Machine (110) with a Q-Metric kernel function computer (112) is provided. The Support Vector Machine (110) exhibits improved performance for classification and regression. Pattern recognition systems (100,900) that use the Support Vector Machine (110) are also provided. A Differential Evolution method of training a Support Vector Machine is also provided.

Abstract: Virtual reality experiences are provided (101 and 102) for a first participant and a second participant. A virtual representation of the second participant's interaction with the shared experience is rendered (103) for the first participant. Similarly, a virtual representation of the second participant's interaction with the shared experience is rendered (104) for the second participant. Upon detecting (105) an interaction between the second participant and a shared virtual component, the virtual representation for the first participant of the second participant's interaction with the shared experience is rendered (106) as though the interaction between the second participant and the shared virtual component had not occurred notwithstanding that the rendering as provided to the second participant does reflect and incorporate that interaction.

Abstract: In a parallel computation of a Hough transform of an array of input data values, the transform space of the Hough transform is partitioned dynamically or statically into a number of sub-spaces. Each sub-space of the transform is stored in a sub-space of memory locations. Data values from the array of input data values are passed to a plurality of processors, each processor associated dynamically or statically with a sub-space of memory locations. Each processor, acting in parallel with the other processors, updates constituent elements of the Hough transform stored in the associated sub-space memory locations dependent upon the input data value.