Abstract: A method, apparatus and system for anomaly detection in a processor based system includes training a deep learning sequence prediction model using observed baseline behavioral sequences of at least one processor behavior of the processor based system, predicting baseline behavioral sequences from the observed baseline behavioral sequences using the sequence prediction model, determining a baseline reconstruction error distribution profile using the baseline behavioral sequences and the predicted baseline behavioral sequences, predicting test behavioral sequences from observed, test behavioral sequences using the sequence prediction model, determining a testing reconstruction error distribution profile using the observed test behavioral sequences and the predicted test behavioral sequences, and comparing the baseline reconstruction error distribution profile to the testing reconstruction error distribution profile to determine if an anomaly exists in a processor behavior of the processor based system.

Abstract: Modules and control units cooperate to simultaneously and independently control and adjust pixel parameters non-uniformly at regional increments across an entire image captured in an image frame by pixels in a pixel array. Pixel parameter changes for pixels in a given region occur, based on i) a contextual understanding of what contextually was happening in the one or more prior image frames and ii) whether salient items are located within that region. Additionally, guidance is sent to the sensor control unit to i) increase or decrease pixel parameters within those regions with salient items and then either to i) maintain, ii) increase or iii) decrease pixel parameters within regions without salient items in order to stay within any i) bandwidth limitations ii) memory storage, and/or iii) power consumptions limitations imposed by 1) one or more image sensors or 2) the communication loop between the sensor control unit and the image processing unit.

Abstract: Operations of computing devices are managed using one or more deep neural networks (DNNs), which may receive, as DNN inputs, data from sensors, instructions executed by processors, and/or outputs of other DNNs. One or more DNNs, which may be generative, can be applied to the DNN inputs to generate DNN outputs based on relationships between DNN inputs. The DNNs may include DNN parameters learned using one or more computing workloads. The DNN outputs may be, for example, control signals for managing operations of computing devices, predictions for use in generating control signals, warnings indicating an acceptable state is predicted, and/or inputs to one or more neural networks. The signals enhance performance, efficiency, and/or security of one or more of the computing devices. DNNs can be dynamically trained to personalize operations by updating DNN weights or other parameters.

Abstract: Modules and control units cooperate to simultaneously and independently control and adjust pixel parameters non-uniformly at regional increments across an entire image captured in an image frame by pixels in a pixel array. Pixel parameter changes for pixels in a given region occur, based on i) a contextual understanding of what contextually was happening in the one or more prior image frames and ii) whether salient items are located within that region. Additionally, guidance is sent to the sensor control unit to i) increase or decrease pixel parameters within those regions with salient items and then either to i) maintain, ii) increase or iii) decrease pixel parameters within regions without salient items in order to stay within any i) bandwidth limitations ii) memory storage, and/or iii) power consumptions limitations imposed by 1) one or more image sensors or 2) the communication loop between the sensor control unit and the image processing unit.

Abstract: A method, apparatus and system for anomaly detection in a processor based system includes training a deep learning sequence prediction model using observed baseline behavioral sequences of at least one processor behavior of the processor based system, predicting baseline behavioral sequences from the observed baseline behavioral sequences using the sequence prediction model, determining a baseline reconstruction error distribution profile using the baseline behavioral sequences and the predicted baseline behavioral sequences, predicting test behavioral sequences from observed, test behavioral sequences using the sequence prediction model, determining a testing reconstruction error distribution profile using the observed test behavioral sequences and the predicted test behavioral sequences, and comparing the baseline reconstruction error distribution profile to the testing reconstruction error distribution profile to determine if an anomaly exists in a processor behavior of the processor based system.

Abstract: Operations of computing devices are managed using one or more deep neural networks (DNNs), which may receive, as DNN inputs, data from sensors, instructions executed by processors, and/or outputs of other DNNs. One or more DNNs, which may be generative, can be applied to the DNN inputs to generate DNN outputs based on relationships between DNN inputs. The DNNs may include DNN parameters learned using one or more computing workloads. The DNN outputs may be, for example, control signals for managing operations of computing devices, predictions for use in generating control signals, warnings indicating an acceptable state is predicted, and/or inputs to one or more neural networks. The signals enhance performance, efficiency, and/or security of one or more of the computing devices. DNNs can be dynamically trained to personalize operations by updating DNN weights or other parameters.

Abstract: A method and apparatus for image processing a lens-distorted image (e.g., a fisheye image) is provided. The method includes partitioning coordinate points in a selected output image into tiles. The output image is an undistorted rendition of a subset of the lens-distorted image. Coordinate points on a border of the tiles in the output image are selected. For each tile, coordinate points in the lens-distorted image corresponding to each selected coordinate point in the output image are calculated. In addition, for each tile, a bounding box on the lens-distorted image is selected. The bounding box includes the calculated coordinates in the lens-distorted image. The bounding boxes are expanded so that they encompass all coordinate points in the lens-distorted image that map to all coordinate points in their respective corresponding tiles. Output pixel values are generated for each tile from pixel values in their corresponding expanded bounding boxes.

Abstract: A multi-media device comprising a controller configured to determining that media presented on a multimedia interface will be subject to starvation based on a rate at which a stream of data is received and a rate at which the media is presented on a multimedia interface of the device. The controller also configured to interrupt the presentation of media on the multimedia interface and to present supplemental content on the multimedia interface starvation occurs, to buffer data from the stream of data while presenting the supplemental content, and to present media on the multimedia interface based at least partly on the buffered data after presentation of the supplemental content.

Abstract: A method and apparatus for image processing a lens-distorted image (e.g., a fisheye image) is provided. The method includes partitioning coordinate points in a selected output image into tiles. The output image is an undistorted rendition of a subset of the lens-distorted image. Coordinate points on a border of the tiles in the output image are selected. For each tile, coordinate points in the lens-distorted image corresponding to each selected coordinate point in the output image are calculated. In addition, for each tile, a bounding box on the lens-distorted image is selected. The bounding box includes the calculated coordinates in the lens-distorted image. The bounding boxes are expanded so that they encompass all coordinate points in the lens-distorted image that map to all coordinate points in their respective corresponding tiles. Output pixel values are generated for each tile from pixel values in their corresponding expanded bounding boxes.

Abstract: An interface (402) to a video imaging device (490) that generates captured frames of a video image (100) is described. The interface includes control registers (430) that store at least two sets of stream parameters (401) corresponding to at least two regions of interest (ROIs) and a function (410) that generates at least two corresponding ROI video streams (460) that are derived from the captured frames as essentially simultaneous output signals using the at least two sets of stream parameters. In some embodiments a new selection value (436) is determined from a current selection value and a derivative parameter stored as a portion of a set of stream parameters. In these embodiments, the generating of the at least two corresponding ROI video streams as essentially simultaneous output signals includes using the new selection value.

Abstract: Described herein are systems and methods for expanding upon the single-distance-based background denotation to seamlessly replace unwanted image information in a captured image derived from an imaging application so as to account for a selected object's spatial orientation to maintain an image of the selected object in the captured image.

Abstract: A specification of a configurable processor is generated by generating (1) specifications of first and second stream memory interfaces to be operable to access data in accordance with first and second stream descriptors, and (2) a specification of an interim data storage device (buffer) to be accessed by the first and second stream memory interfaces and to be operable to receive data from a first computational module via the first stream memory interface and to transfer data to a second computational module via the second stream memory interface. The specifications are output and may be used to configure a configurable processor.

Abstract: A method and apparatus for image processing a lens-distorted image (e.g., a fisheye image) is provided. The method includes partitioning coordinate points in a selected output image into tiles. The output image is an undistorted rendition of a subset of the lens-distorted image. Coordinate points on a border of the tiles in the output image are selected. For each tile, coordinate points in the lens-distorted image corresponding to each selected coordinate point in the output image are calculated. In addition, for each tile, a bounding box on the lens-distorted image is selected. The bounding box includes the calculated coordinates in the lens-distorted image. The bounding boxes are expanded so that they encompass all coordinate points in the lens-distorted image that map to all coordinate points in their respective corresponding tiles. Output pixel values are generated for each tile from pixel values in their corresponding expanded bounding boxes.

Abstract: Memory addresses for a data stream are generated by a stream parameter generator that calculates a set of stream parameters for each of a number of memory access patterns and a regional address generator that calculates a sequence of addresses of a memory access pattern from a corresponding set of stream parameters. The stream parameters, which may include START_ADDRESS, STRIDE, SKIP and SPAN values for example, are updated in accordance with an update( ) function. The update( ) function, which may be defined by a user, defines how stream parameters change from one memory access pattern to the next. In one application, the update( ) function describes how the position, shape and/or size of a region of interest in an image changes or is expected to change.

Abstract: Described herein are systems and methods for expanding upon the single-distance-based background denotation to seamlessly replace unwanted image information in a captured image derived from an imaging application so as to account for a selected object's spatial orientation to maintain an image of the selected object in the captured image.

Abstract: A streaming data interface device (700) of a streaming processing system (200) is automatically generated by selecting a set of circuit parameters (610) consistent with a set of circuit constraints and generating (612, 614) a representation of a candidate memory interface device based upon a set of stream descriptors. The candidate streaming data interface device is evaluated (616) with respect to one or more quality metrics and the representation of the candidate streaming processor circuit is output (622) if the candidate memory interface device satisfies a set of processing system constraints and is better in at least one of the one or more quality metrics than other candidate memory interface devices.

Abstract: An alert system is activated in response to an event by positioning a portable electronic device, a cellular telephone or PDA for example, such that a camera of the electronic device is orientated to capture images relating to a visual criterion by which the event is to be detected. Images relating to the visual criterion are processed to detect the event and an alert system is activated if an event is detected. In one application, the portable device is held in a holder in a vehicle and used to detect vehicle lane departure or driver drowsiness.

Abstract: A content addressable memory (CAM) is disclosed that includes a memory having a first port configured to write a 1-bit data to the memory and a second port configured to read and write N-bit data. To update the CAM, an N-bit zero data word is written to the second port at a first address A2 to erase N bits of the memory at address A2, then the first address A2 is combined with a data value A0 to form a second address A1. Finally, a value 1 is written to the first data port at the address A1. Reading the second port at the first address A2 will produce an N-bit data word having value 1 at bit position A0 and zeros at all other bit positions. The CAM may be configured in reconfigurable hardware using random access memory and used in a stream data interface circuit.

Abstract: A specification of a configurable processor is generated by generating (1) specifications of first and second stream memory interfaces to be operable to access data in accordance with first and second stream descriptors, and (2) a specification of an interim data storage device (buffer) to be accessed by the first and second stream memory interfaces and to be operable to receive data from a first computational module via the first stream memory interface and to transfer data to a second computational module via the second stream memory interface. The specifications are output and may be used to configure a configurable processor.

Abstract: A content addressable memory (CAM) is disclosed that includes a memory having a first port configured to write a 1-bit data to the memory and a second port configured to read and write N-bit data. To update the CAM, an N-bit zero data word is written to the second port at a first address A2 to erase N bits of the memory at address A2, then the first address A2 is combined with a data value A0 to form a second address A1. Finally, a value 1 is written to the first data port at the address A1. Reading the second port at the first address A2 will produce an N-bit data word having value 1 at bit position A0 and zeros at all other bit positions. The CAM may be configured in reconfigurable hardware using random access memory and used in a stream data interface circuit.