Abstract: In accordance with the principles herein, a co-design approach for compute-in-memory inference for deep neural networks (DNN) is set forth. Multiplication-free function approximators are employed along with a co-adapted processing array and compute flow. Resulting methods, systems, devices, and algorithms in accordance with the principles herein overcome many deficiencies in the currently available in—methods, systems, devices, and algorithms (in-SRAM) DNN processing devices. Systems, devices, and algorithms constructed in accordance with the co-adapted implementation herein seamlessly extends to multi-bit precision weights, eliminates the need for DACs, and easily extends to higher vector-scale parallelism. Additionally, a SRAM-immersed successive approximation ADC (SA-ADC) can be constructed, where the parasitic capacitance of bit lines of SRAM array can be exploited as a capacitive DAC. The dominant area overhead in SA-ADC, due to its capacitive DAC, can allow low area implementation of within-SRAM SA-ADC.

Abstract: Methods and systems for monitoring a subject in a sleep or resting state are provided herein. The methods can include obtaining signals from at least one of one or more microphones, one or more pyroelectric infrared (PIR) sensors and one or more accelerometer sensors. The obtained signals can be automatically transmitted to a processor. Thereafter, using the processor, one or more patterns in the obtained signals can be detected to determine one or more physiological and/or biological parameters including at least one of heart rate, breathing rate, wheezing, sleep quality and/or sleep architecture. Outputs of the determined parameters may be generated upon crossing preset thresholds.

Abstract: Wildfires are detected by controlling image scanning within the viewing range of a video camera to generate digital images that are analyzed to detect gray colored regions, and then to determine whether a detected gray colored region is smooth. Further analysis to determine movement in a gray colored smooth region uses a past image which is within a slow moving time range, as determined by a strategy for controlling the image scanning. Additional analysis connects a candidate region to a land portion of the image, and a support vector machine is applied to a covariance matrix of the candidate region to determine whether the region shows smoke from a wildfire.

Abstract: Wildfires are detected by controlling image scanning within the viewing range of a video camera to generate digital images that are analyzed to detect gray colored regions, and then to determine whether a detected gray colored region is smooth. Further analysis to determine movement in a gray colored smooth region uses a past image which is within a slow moving time range, as determined by a strategy for controlling the image scanning. Additional analysis connects a candidate region to a land portion of the image, and a support vector machine is applied to a covariance matrix of the candidate region to determine whether the region shows smoke from a wildfire.

Abstract: A video based method to detect volatile organic compounds (VOC) leaking out of components used in chemical processes in petrochemical refineries. Leaking VOC plume from a damaged component has distinctive properties that can be detected in realtime by an analysis of images from a combination of infrared and optical cameras. Particular VOC vapors have unique absorption bands, which allow these vapors to be detected and distinguished. A method of comparative analysis of images from a suitable combination of cameras, each covering a range in the IR or visible spectrum, is described. VOC vapors also cause the edges present in image frames to loose their sharpness, leading to a decrease in the high frequency content of the image. Analysis of image sequence frequency data from visible and infrared cameras enable detection of VOC plumes. Analysis techniques using adaptive background subtraction, sub-band analysis, threshold adaptation, and Markov modeling are described.

Abstract: The present invention describes a method and a system for detection of fire and smoke using image and video analysis techniques to detect the presence of indicators of fire and smoke. The method and the system detects smoke by transforming plurality of images forming the video captured by a camera into Nonlinear Median filter Transform (NMT) domain, implementing an “L1”-norm based energy measure indicating the existence of smoke from the MMT domain data, detecting slowly decaying NMT coefficients, performing color analysis in low-resolution NMT sub-images, using a Markov model based decision engine to model the turbulent behavior of smoke, and fusing the above information to reach a final decision about the existence of smoke within the viewing range of camera.

Abstract: A video based method to detect volatile organic compounds (VOC) leaking out of components used in chemical processes in petrochemical refineries. Leaking VOC plume from a damaged component causes edges present in image frames to loose their sharpness, leading to a decrease in the high frequency content of the image. Analysis of image sequence frequency data from visible and infrared cameras enable detection of VOC plumes in real-time. Analysis techniques using adaptive background subtraction, sub-band analysis, threshold adaptation, and Markov modeling are described.

Abstract: A system for automatically expanding the zoom capability of a wide-angle video camera using images from multiple camera locations. One preferred embodiment achieves this using images from the wide-angle video camera that are analyzed to identify regions of interest (RoI). Pan-Tilt-Zoom (PTZ) controls are then sent to aim slave cameras toward the RoI. Processing circuitry is then used to replace the RoI from the wide-angle images with the higher-resolution images from one of the slave cameras. In addition, motion-detecting software can be utilized to automatically detect, track, and/or zoom in on moving objects.

Abstract: An enhanced infrared (IR) imaging based method for detecting volatile organic and hazardous vapors using the infrared spectral absorption properties of these vapors, and using a tunable infrared light source and a plurality of cameras tuned to particular frequency ranges to detect spectral absorption properties corresponding to the respective vapors. Illumination by an IR light source is used to enhance the visibility of vapor plumes in LWIR and MWIR cameras because ambient light may not have enough power in the specific absorption band of the VOC vapor. Plume regions are automatically determined by image and video processing methods by the system. Specific vapors can be detected by using tunable IR light sources because leaking plumes from a damaged component causes dark regions in images of LWIR and/or MWIR cameras depending on the absorption wavelength of the plume.

Abstract: A system for automatically expanding the zoom capability of a wide-angle video camera using images from multiple camera locations. One preferred embodiment achieves this using images from the wide-angle video camera that are analyzed to identify regions of interest (RoI). Pan-Tilt-Zoom (PTZ) controls are then sent to aim slave cameras toward the RoI. Processing circuitry is then used to replace the RoI from the wide-angle images with the higher-resolution images from one of the slave cameras. In addition, motion-detecting software can be utilized to automatically detect, track, and/or zoom in on moving objects.

Abstract: A system for automatically expanding the zoom capability of a wide-angle video camera using images from multiple camera locations. One preferred embodiment achieves this using images from the wide-angle video camera that are analyzed to identify regions of interest (RoI). Pan-Tilt-Zoom (PTZ) controls are then sent to aim slave cameras toward the RoI. Processing circuitry is then used to replace the RoI from the wide-angle images with the higher-resolution images from one of the slave cameras. In addition, motion-detecting software can be utilized to automatically detect, track, and/or zoom in on moving objects.

Abstract: A system and method for varying the level of detail encoded in a video. In a preferred embodiment, some regions of a wide-angle video scene are encoded in an almost lossless manner, while other regions are encoded with less detail. Regions of interest can be determined in many ways, including a priori, automatically in real time, or by the selection of human operators.

Abstract: A video based method to detect volatile organic compounds (VOC) leaking out of components used in chemical processes in petrochemical refineries. Leaking VOC plume from a damaged component causes edges present in image frames to loose their sharpness, leading to a decrease in the high frequency content of the image. Analysis of image sequence frequency data from visible and infrared cameras enable detection of VOC plumes in real-time. Analysis techniques using adaptive background subtraction, sub-band analysis, threshold adaptation, and Markov modeling are described.

Abstract: A method and system for characterizing the motion of moving objects in video is disclosed. The video can be either compressed or uncompressed form. A typical video scene contains foreground and background objects. Foreground objects temporarily stay in the video. However, a left object or stopped object becomes part of the background scene and remains in the video forever. A plurality of images is inputted to the system in time series. A method and system determines left objects or stopped objects by comparing the background image estimated from the current image of the video with the background estimated from previous images of the video. Difference between the current and previous background images indicates a left or stopped object. Other objects, which do not modify the background image are determined as transitory objects. Background scene of a video can be estimated using the compressed video data as well.