Abstract: Conventional gesture detection approaches demand large memory and computation power to run efficiently, thus limiting their use in power and memory constrained edge devices. Present application/disclosure provides a Spiking Neural Network based system which is a robust low power edge compatible ultrasound-based gesture detection system. The system uses a plurality of speakers and microphones that mimics a Multi Input Multi Output (MIMO) setup thus providing requisite diversity to effectively address fading. The system also makes use of distinctive Channel Impulse Response (CIR) estimated by imposing sparsity prior for robust gesture detection. A multi-layer Convolutional Neural Network (CNN) has been trained on these distinctive CIR images and the trained CNN model is converted into an equivalent Spiking Neural Network (SNN) via an ANN (Artificial Neural Network)-to-SNN conversion mechanism. The SNN is further configured to detect/classify gestures performed by user(s).

Abstract: This disclosure relates generally to Synthetic Aperture Radar (SAR) reconstruction and finds wide application in remote sensing. Conventional approaches either involve huge computational requirement for processing or require specialized hardware along with many additional Radio Frequency (RF) components. The present disclosure provides two approaches for temporally sampling a received pulse compressed signal at two sub-sampling factors, wherein both methods involve frugal hardware implementation. Reconstruction approach of the art is based on the principle of difference ruler and is not suitable for SAR image reconstruction due to the large measurements and image dimensions. In accordance with the present disclosure, the reconstruction problem is framed as an inverse imaging problem by suitably using a forward model and employing an approach like Alternating Direction Method of Multipliers (ADMM) for solving this model which allows use of readily available Plug and Play (PnP) priors.

Abstract: The present invention relates to a method and system for Phaseless Passive Synthetic Aperture Radar (PPSAR) imaging. Existing method for image reconstruction requires large number of measurements for satisfactory PPSAR image reconstruction. However, this leads to provisioning of more on-board storage and/or a high-speed data link between a mobile platform and a ground station. These requirements are undesirable in practice as PPSAR image reconstruction systems are deployed on resource constrained platforms. The present disclosure uses a regularized Wirtinger Flow (rWF) based approach that uses appropriate regularizers to facilitate the PPSAR image reconstruction with fewer measurements. Further the PPSAR image reconstruction is achieved using Alternating Direction Method of Multipliers (ADMM) by employing standard denoisers such as Total Variation (TV), Block-matching and 3D filtering (BM3D) and, Deep Image Prior (DIP).

Abstract: This disclosure relates to a method and system for multi-sensor fusion in the presence of missing and noisy labels. Prior methods for multi-sensor fusion do not estimate and correct labels for learning effective models in semi-supervised learning methods. Embodiments of the present disclosure provides a method for learning robust sensor-specific autoencoder based fusion model by utilizing a graph structure to perform label propagation and correction. In the disclosed Graph regularized AutoFuse (GAF) method latent representation for each sensor is learnt using the sensor-specific autoencoders. Further these latent representations are combined and fed to a classifier for multi-class classification. The disclosure presents a joint optimization formulation for multi-sensor fusion where label propagation and correction, sensor-specific learning and classification are executed together.

Abstract: This disclosure relates generally to material quality inspection. Conventional approaches available for material quality inspection are unable to address concerns of complexity and cost involved. The technical problem of occluded object detection and material quality inspection for intrinsic defects identification is addressed in the present disclosure. The present disclosure provides a system and method for non-intrusive material quality inspection using three-dimensional monostatic radar based imaging, where the object under inspection undergoes a circular translation motion on a rotating platform. A modified delay-and-sum (m-DAS) algorithm is built by incorporating virtual antenna array to obtain a 3D image reconstruction of the object. From 3D reconstructed images, radial displacement as well as the angular locations of the object is identified which are further used for quality inspection of the material comprised in the object.

Abstract: Conventional gesture detection approaches demand large memory and computation power to run efficiently, thus limiting their use in power and memory constrained edge devices. Present application/disclosure provides a Spiking Neural Network based system which is a robust low power edge compatible ultrasound-based gesture detection system. The system uses a plurality of speakers and microphones that mimics a Multi Input Multi Output (MIMO) setup thus providing requisite diversity to effectively address fading. The system also makes use of distinctive Channel Impulse Response (CIR) estimated by imposing sparsity prior for robust gesture detection. A multi-layer Convolutional Neural Network (CNN) has been trained on these distinctive CIR images and the trained CNN model is converted into an equivalent Spiking Neural Network (SNN) via an ANN (Artificial Neural Network)-to-SNN conversion mechanism. The SNN is further configured to detect/classify gestures performed by user(s).

Abstract: This relates generally to a method and a system for spatio-temporal polarization video analysis. The spatio-temporal polarization data is analyzed for a computer vision application such as object detection, image classification, image captioning, image reconstruction or image inpainting, face recognition and action recognition. Numerous classical and deep learning methods have been applied on polarimetric data for polarimetric imaging analysis, however, the available pre-trained models may not be directly suitable on polarization data, as polarimetric data is more complex. Further compared to analysis of the polarimetric images, a significant number of actions can be detected by polarimetric videos, hence analyzing polarimetric videos is more efficient. The disclosure is a spatio-temporal analysis of polarization video.

Abstract: Existing Mueller Matrix polarization techniques that rely only on polarization properties are insufficient for accurate characterization of transparent objects. Embodiments of the present disclosure provide a method and system for Mueller Matrix polarimetric characterization of transparent object using optical properties along with the polarization properties to accurately characterize the transparent object. The polarization properties of are derived from a decomposed Mueller matrix element. Additionally, the method derives the optical properties by further subjecting the decomposed Mueller matrix element to Fresnel’s law-based analysis and a reverse Monte Carlo analysis to extract optical properties such as a material refractive index and a material attenuation index. Optical properties vary with changes in the material property caused due to several factors such as manufacturing defect, aberration, inclusion of an impurity such as bubble or dust etc.

Abstract: Gesture recognition is a key requirement for Human Computer Interaction (HCI) and multiple modalities are explored in literature. Conventionally, channel taps are estimated using least square based estimation and tap corresponding to finger motion is tracked. These assume that noise component is negligible and can reduce the tracking accuracy for low SNR. Thus, to mitigate the above-mentioned limitation, the system and method of the present disclosure explore the feasibility of using speaker and microphone setup available in most of smart devices and transmit inaudible frequencies (acoustic) for detecting the human finger level gestures accurately. More specifically, System implements the method for millimeter level finger tracking and low power gesture detection on this tracked gesture. The system uses a subspace based high resolution technique for delay estimation and use microphone pairs to jointly estimate the multi-coordinates of finger movement.

Abstract: Use of Swept Frequency Acoustic Interferometry (SFAI) is becoming ubiquitous in taking non-invasive measurements of fluid parameters like sound speed, sound attenuation and density of fluid. But measurement using SFAI is relatively slow as one needs to sweep a wide range of frequencies and for each probing frequency one needs to wait for settling time. Further, SFAI works well only on steady flow as sudden change in fluid flow destroys resonance condition, thereby making it unsuitable for flowing fluid. Present application provides method and system for faster assessment of sound speed in fluids using compressive sensing technique. The system first uses random samples in defined frequency scanning range of frequency sweep signal for generating pseudo analytic signal vector. The system then estimates pulse-echo view by applying compressive sensing technique over pseudo analytic signal vector.

Abstract: This disclosure relates generally to a method and system for multi-modal image super-resolution. Conventional methods for multi-modal image super-resolution are performed using joint image based filtering, deep learning and dictionary based approaches which require large datasets for training. Embodiments of the present disclosure provide a joint optimization based transform learning framework wherein a high-resolution (HR) image of target modality is reconstructed from a HR image of guidance modality and a low-resolution (LR) image of target modality. A set of parameters, transforms, coefficients and weight matrices are learnt jointly from a training data which includes a HR image of guidance modality, a LR image of target modality and a HR image of target modality. The learnt set of parameters are used for reconstructing a HR image of target modality. The disclosed joint optimization transform learning framework is used in remote sensing, environment monitoring and so on.

Abstract: This disclosure relates generally to Millimeter Wave (MMW) frequency antenna scanning system. Conventional approaches available for scanning an antenna beam over a large angular swath with high directivity are unable to address concerns of size and cost involved. The technical problem of providing an MMW frequency antenna scanning system using a single small size antenna capable of scanning as desired at a desired precision is addressed in the present disclosure. The antenna scanning system provided is an electromechanical system that makes the system cost effective. Computer control provides precision control in beam steering from remote. Use of a metasurface and configuration of a radiating patch and a shorting pin in a microstrip antenna addresses the concern with regards to the size of the antenna scanning system.

Abstract: Systems and methods for a dense surface reconstruction of an object using graph signal processing is provided. None of the traditional systems and methods provide for a dense or three-dimensional surface reconstruction of objects by resolving ? ambiguity. The embodiments of the proposed disclosure provide for resolving ? ambiguity by identifying, from one or more sparse three-dimensional shapes extracted, a first set of azimuth values corresponding to a first region of the object; constructing, using a phase angle, a graph capturing a relational structure between the first set of azimuth values and a second set of azimuth values to be estimated; obtaining, a Graph Fourier Transform (GFT) matrix corresponding to the constructed graph; and estimating, from the GFT matrix and the first set of azimuth values, the second set of azimuth values corresponding to a second region of the object by the graph signal processing technique.

Abstract: This disclosure relates to multi-sensor fusion using Transform Learning (TL) that provides a compact representation of data in many scenarios as compared to Dictionary Learning (DL) and Deep network models that may be computationally intensive and complex. A two-stage approach for better modeling of sensor data is provided, wherein in the first stage, representation of the individual sensor time series is learnt using dedicated transforms and their associated coefficients and in the second stage, all the representations are fused together using a fusing (common) transform and its associated coefficients to effectively capture correlation between the different sensor representations for deriving an inference. The method and system of the present disclosure can find application in areas employing multiple sensors that are mostly heterogeneous in nature.

Abstract: The disclosure herein generally relates to image processing, and, more particularly, to a method and system for impurity detection using multi-modal image processing. This system uses a combination of polarization data, and at least one of a depth data and an RGB image data to perform the impurity material detection. The system uses a graph fusion based approach while processing the captured images to detect presence of the impurity material, and accordingly alert the user.

Abstract: This disclosure relates generally to a method and system for assessment of sustained visual attention of a target. The conventional methods utilize various markers for assessment of attention, however, blink rate variability (BRV) series signal is not explored yet. In an embodiment, the disclosed method utilizes BRV series signal for assessing sustained visual attention of a target. A gaze data of the target is recorded using an eye tracker and a set of uniformly sampled BRV series signal is reconstructed from each of the BRV series. One or more frequency domain features, including pareto frequency, are extracted from the set of uniformly sampled BRV series signal. The values of frequency domain features extracted from the set of BRV series signals are compared with corresponding threshold values to determine visual sustained attention of the target.

Abstract: This disclosure relates to attention shifting of a robot in a group conversation with two or more attendees, wherein at least one of them is a speaker. State of the art has dealt with several aspects of Human-Robot Interaction (HRI) including responding to a source of sound at a time, addressing a fixed viewing area or determining who is the speaker based on eye gaze direction. However, attention shifting to make the conversation human-like is a challenge. The present disclosure uses audio-visual perception for speaker localization. Only qualified direction of arrivals (DOAs) are used for the audio perception. Further the audio perception is complimented by visual perception employing real time face detection and lip movement detection. Use of HRI rules, clustering of the DOAs, dynamic adjustment of rotation of the robot and a dynamically updated knowledge repository enriches the robot with intelligence to shift attention with minimum human intervention.

Abstract: This disclosure relates generally to method and system for low sampling rate electrical load disaggregation. At low sampling rates, disaggregation of energy load is challenging due to unavailability of events and signatures of the constituent loads. The disclosed energy disaggregation technique receives aggregated load data from a utility meter and sequentially obtains training data for determining disaggregated energy load at low sampling rate. Dictionaries are used to characterize the different loads in terms of power values and time of operation. The obtained dictionary coefficients are treated as graph signals and graph smoothness is used for propagating the coefficients from the training phase to the test phase by formulating an optimization model. The derivation of the optimization model identifies the load of interest and estimate their power consumption based on optimization model constraints. This method achieves accuracy greater than 70% for the loads of interest at low sampling rates.

Abstract: This disclosure relates generally to field agnostic source localization. Conventional state-of-the-art methods perform source localization for near-field scenario by estimating carrier frequency and direction of arrival (DOA) at or above Nyquist sampling rate. Embodiments of the present disclosure provide a method for source localization at sub Nyquist sampling rate. The method estimates parameters such as range, carrier frequency and DOA of source signals from data sources in a mixed field scenario. i.e., the data sources may reside in far-field as well as near-field. The method considers a delay channel to a sensor receiver architecture for estimating the parameters. The disclosed method can be used in applications like cognitive radio to determine the carrier frequency, DOA and range of various source signals from data sources in mixed field.

Abstract: This disclosure relates generally to field agnostic source localization. Conventional state-of-the-art methods perform source localization for near-field scenario by estimating carrier frequency and direction of arrival (DOA) at or above Nyquist sampling rate. Embodiments of the present disclosure provide a method for source localization at sub Nyquist sampling rate. The method estimates parameters such as range, carrier frequency and DOA of source signals from data sources in a mixed field scenario. i.e., the data sources may reside in far-field as well as near-field. The method considers a delay channel to a sensor receiver architecture for estimating the parameters. The disclosed method can be used in applications like cognitive radio to determine the carrier frequency, DOA and range of various source signals from data sources in mixed field.