Abstract: Systems, devices and methods are provided for building learning machines using learning machines. The system generally includes a reference learning machine, a target learning machine being built, a component analyzer module configured to analyze inputs from the reference learning machine, the target learning machine, a set of test signals, and a list of components in the reference learning machine and the target learning machine, and return a set of output values for each component on the list of components. The system further includes a component tuner module configured to modify different components in the target learning machine based on the set of output values and a component mapping, thereby resulting in a tuned learning machine.

Abstract: A method for automatically computing the position and score for each dart on a dartboard using a single image of the dart and dartboard captured from any camera angle. A computing device with at least one connected camera captures an image of a dartboard. The computing device performs various image processing tasks and computes the landing position of each dart present in the image and the positions of multiple dartboard calibration points. The dartboard calibration points are used to compute a transformation matrix that transforms the detected points in the image plane to corresponding points in the dartboard plane. The score of each dart is computed based on the relative positioning of the transformed points in the dartboard plane. The method may be used to perform automatic scorekeeping in a variety of different dart games, either in a single location or in remote locations.

Abstract: Systems, devices and methods are provided for building learning machines using learning machines. The system generally includes a reference learning machine, a target learning machine being built, a component analyzer module configured to analyze inputs from the reference learning machine, the target learning machine, a set of test signals, and a list of components in the reference learning machine and the target learning machine, and return a set of output values for each component on the list of components. The system further includes a component tuner module configured to modify different components in the target learning machine based on the set of output values and a component mapping, thereby resulting in a tuned learning machine.

Abstract: There is disclosed a novel system and method for decentralized digital structured data storage, management, and authentication. In an embodiment, the present system comprises pieces of digital structured data (including but not limited to, digital images, digital videos, digital audio, digital text, and digital computational graph representations), computer nodes for storing pieces of digital structured data and for creating references to portions of stored pieces of digital structured data, a network connecting the computer nodes, and a blockchain. In an embodiment, a piece of digital structured data is stored on a computer node, and the computer node creates a reference to a portion of the stored piece of digital structured data. The computer node then broadcasts a request for the reference to be added to the blockchain to all computer nodes through a network connecting all computer nodes.

Abstract: There is disclosed a novel system and method for decentralized digital structured data storage, management, and authentication. In an embodiment, the present system comprises pieces of digital structured data (including but not limited to, digital images, digital videos, digital audio, digital text, and digital computational graph representations), computer nodes for storing pieces of digital structured data and for creating references to portions of stored pieces of digital structured data, a network connecting the computer nodes, and a blockchain. In an embodiment, a piece of digital structured data is stored on a computer node, and the computer node creates a reference to a portion of the stored piece of digital structured data. The computer node then broadcasts a request for the reference to be added to the blockchain to all computer nodes through a network connecting all computer nodes.

Abstract: An automatic system and method for non-invasive imaging and identification of specific ocular structures of the eye and adnexa tissues by synchronous segmentation of visual and infrared images; that can produce spatial temperature profiles within each segmented area of the eye and adnexa; that can track eye and head movement and eye-blinks during the period of measurement to remove artefacts and maintain synchronicity; that can track ocular surface and eye adnexa temperature profiles over time; that can assist in diagnosis of eye disease; that can produce diagnostic indicators for ocular disease diagnosis and study of the eye. The system comprises infrared and visible light cameras for imaging the ocular structures, and a digital signal processing unit for processing the acquired infrared and visible images to output segmentations of the images for identification of different areas of the eye surface, including pupil, cornea, conjunctiva, and eyelids.

Abstract: Systems, devices and methods are provided for building and using learning machines to understand and explain learning machines. The present system comprises a reference learning machine and an explainer learning machine being built for explaining and understanding the reference learning machine. A set of input signals is fed through the reference learning machine and the outputs at the different components of the learning machine for each given input signal are recorded. The recorded outputs at the different components of the learning machine for each given input signal, along with the corresponding expected output of the learning machine for each given input signal, are then used to update the parameters of the explainer learning machine. After the parameter update process, the explainer learning machine can then be queried for quantitative insights about the reference learning machine.

Abstract: A method and apparatus for predicting hospital bed exit events from video camera systems is disclosed. The system processes video data with a deep convolutional neural network consisting of five main layers: a 1×1 3D convolutional layer used for generating feature maps from raw video data, a context-aware pooling layer used for rectifying data from different camera angles, two fully connected layers used for applying pre-trained deep features, and an output layer used to provide a likelihood of a bed exit event.

Abstract: There is disclosed a system and method for measuring the polarization of light which utilizes a spatially-varying polarization beam. This measurement is analyzed and improved using novel imaging methods and image processing methods on a digital signal processing (DSP) system to determine the polarization of light. Various embodiments are described which can measure light polarization, determine polarizing material distributions, determine the optical rotary dispersion of samples, and determine the circular dichroism of samples. In certain embodiments, the binding constants and binding activity of molecular samples can be achieved. Given its reduced complexity and size, the system may be configured to be portable such that the system may be used in various applications outside of a lab setting.

Abstract: There is disclosed a system and method for measuring arterial and venous blood pulse waveforms (BPWs) of a subject utilizing photoplethysmography (PPG). In an embodiment, the system and method comprises: providing a plurality of virtual sensors positioned to cover a desired field-of-view of the subject, each virtual sensor adapted to detect and measure a BPW signal from an area of the subject's body and provide a BPW signal output; processing the BPW signal outputs of the plurality of virtual sensors to compare the BPWs at multiple areas of the subject's body to perform spatial perfusion analysis; and displaying at least one aggregate output based on the spatial perfusion analysis. At least one aggregate output may include a visualization of one or more perfusion patterns overlaid on a photographic image of the subject, and aggregate statistics including subject heart rate and breathing rate.

Abstract: There is disclosed a novel system and method for multiview, multispectral, polarimetric, light-field, and high dynamic range imaging in a concurrent manner specifically capturing information at different spectral bands and light polarizations simultaneously. The present system and method is capable of (1) concurrent imaging of multiple spectral bands (including spectral bands beyond the visible region of the electromagnetic spectrum), proportional or greater than the number of filters used in the device, (2) concurrent imaging of multiple light polarizations (Stokes vectors), (3) acquiring images at different point-of-view of the same scene and/or object that allow for topographical reconstruction, (4) concurrent imaging of the multiple depth of fields that allow for light-field imaging, and (5) concurrent imaging of multiple simulated exposures of the detector that allow for high dynamic range imaging, all at the same time using a single sensor in the same imaging system enclosure.

Abstract: There is disclosed a novel system and method for decentralized digital structured data storage, management, and authentication. In an embodiment, the present system comprises pieces of digital structured data (including but not limited to, digital images, digital videos, digital audio, digital text, and digital computational graph representations), computer nodes for storing pieces of digital structured data and for creating references to portions of stored pieces of digital structured data, a network connecting the computer nodes, and a blockchain. In an embodiment, a piece of digital structured data is stored on a computer node, and the computer node creates a reference to a portion of the stored piece of digital structured data. The computer node then broadcasts a request for the reference to be added to the blockchain to all computer nodes through a network connecting all computer nodes.

Abstract: There is disclosed a system and method for measuring the polarization of light which utilizes a spatially-varying polarization beam. This measurement is analyzed and improved using novel imaging methods and image processing methods on a digital signal processing (DSP) system to determine the polarization of light. Various embodiments are described which can measure light polarization, determine polarizing material distributions, determine the optical rotary dispersion of samples, and determine the circular dichroism of samples. In certain embodiments, the binding constants and binding activity of molecular samples can be achieved. Given its reduced complexity and size, the system may be configured to be portable such that the system may be used in various applications outside of a lab setting.

Abstract: A method for monitoring a patient in a bed may involve capturing images of the patient with multiple cameras in a vicinity of the bed, wirelessly transmitting the images of the patient from the multiple cameras to a processor including a memory device, processing the images to provide processed image data pertaining to a position of the patient relative to the bed to a user, and analyzing the processed image data to determine whether the patient is exiting the bed. The method may also optionally involve providing an alarm indicating that the patient is exiting the bed.

Abstract: Systems, devices and methods are provided for building learning machines using learning machines. The system generally includes a reference learning machine, a target learning machine being built, a component analyzer module configured to analyze inputs from the reference learning machine, the target learning machine, a set of test signals, and a list of components in the reference learning machine and the target learning machine, and return a set of output values for each component on the list of components. The system further includes a component tuner module configured to modify different components in the target learning machine based on the set of output values and a component mapping, thereby resulting in a tuned learning machine.

Abstract: There is disclosed a novel system and method for achieving ultra-resolution, ultra-wide field-of-view multispectral and hyperspectral holographic microscopy and quantitative phase contrast microscopy. In an embodiment, the method comprises: providing a stationary illumination source; acquiring a plurality of low-resolution holograms of an image subject from different locations utilizing a subpixel sensor-scanning synthetic aperture mechanism whereby a detector scanning translationally, radially and/or rotationally; processing the acquired holograms utilizing a processing algorithm corresponding to the scanning motion of the detector used to acquire the holograms; and reconstructing a subpixel ultra-resolution image of the image subject based on the processed holograms; whereby, a desired synthetic aperture is achieved without loss of resolution. The multispectral and hyperspectral aspect is achieved in the novel system and method by use of different combination of illumination sources (i.e.

Abstract: A method and apparatus for predicting hospital bed exit events from video camera systems is disclosed. The system processes video data with a deep convolutional neural network consisting of five main layers: a 1×1 3D convolutional layer used for generating feature maps from raw video data, a context-aware pooling layer used for rectifying data from different camera angles, two fully connected layers used for applying pre-trained deep features, and an output layer used to provide a likelihood of a bed exit event.

Abstract: A method for monitoring a patient in a bed may involve capturing images of the patient with multiple cameras in a vicinity of the bed, wirelessly transmitting the images of the patient from the multiple cameras to a processor including a memory device, processing the images to provide processed image data pertaining to a position of the patient relative to the bed to a user, and analyzing the processed image data to determine whether the patient is exiting the bed. The method may also optionally involve providing an alarm indicating that the patient is exiting the bed.

Abstract: There is disclosed a novel system and method for improving spectral resolution and signal-to-noise ratio of a captured spectrum by employing digital beam reforming, digital slicing, and digital multiplexing. In an embodiment, the method comprises: and entrance aperture for the light to enter the apparatus, a collimating element (refractive or reflective), a dispersive element (single- or multi-axis), a focusing element (refractive or reflective), and a linear or array detector for acquisition of the spectrum. The spectrum is then digitally beam reformed, digitally beam sliced to increase spectral resolution, and digitally beam multiplexed to increase the signal-to-noise ratio. In addition the disclosed invention is also capable of performing digital beam refocusing which means the optical focusing power can be chosen such that its performance surpasses the performance of any analog optical focusing element thereby further increasing the spectral resolution and the signal-to-noise ratio of the spectrum.

Abstract: There is disclosed a novel system and method for building artificial neural networks for a given task. In an embodiment, the method utilizes one or more network models that define the probabilities of nodes and/or interconnects, and/or the probabilities of groups of nodes and/or interconnects, from sets of possible nodes and interconnects existing in a given artificial neural network. These network models can be constructed based on the properties of one or more artificial neural networks, or constructed based on desired architecture properties. These network models are then used to build combined network models using a model combiner module. The combined network models and random numbers generated by a random number generator module are then used to build one or more new artificial neural network architectures. New artificial neural networks are then built based on the newly built artificial neural network architectures and are trained for a given task.