Tata Consultancy Services Limited

Indian Statistical Institute

Arpan Pal

Soma Bandyopadhyay

Tapas Chakravarty

Aniruddha Sinha

Avik Ghose

Arijit Ukil

Balamuralidhar Purushothaman

Anirban Dutta Choudhury

Chirabrata Bhaumik

Rohan Banerjee

Arijit Sinharay

Arijit Mukherjee

Chetanya Puri

Tanushyam Chattopadhyay

Arijit Chowdhury

Abhijan Bhattacharyya

Shreyasi Datta

Rituraj Singh

Sounak Dey

Swarnava Dey

Achanna Anil Kumar

Brojeshwar Bhowmick

Hemant Kumar Rath

Kingshuk Chakravarty

Gautam Shroff

Puneet Kumar Agarwal

Kayapanda Muthana Mandana

Parijat Dilip Deshpande

Smriti Rani

Dibyanshu Jaiswal

Ramesh Kumar Ramakrishnan

Jayavardhana Rama Gubbi Lakshminarasimha

Chayan Sarkar

Sakyajit Bhattacharya

Prateep Misra

Ranjan Dasgupta

Anwesha Khasnobish

Rajat Kumar Das

Himadri Sekhar Paul

Tanuka Bhattacharjee

Deepan Das

Debatri Chatterjee

Snehasis Banerjee

Vivek Chandel

Dibyendu Roy

Sundeep Khandelwal

Ramu Reddy Vempada

C A Murthy

Rohit Gupta

Rahul Dasharath Gavas

Ashis Sau

Arun George

Kartik Muralidharan

Andrew Gigie

Arindam Saha

Nasimuddin Ahmed

Ayan Mukherjee

Jyotirmoy Karjee

Priyanka Sinha

Hrishav Bakul Barua

Vartika Sengar

Deepak N Swamy

Sanjay Madhukar Kimbahune

Bighnaraj Panigrahi

Pradip Pramanick

Ajay Kattepur

Samar Shailendra

Sujit Shinde

Raj Kumar Rakshit

Prasanta Kumar Ghosh

Amit Swain

Swagata Biswas

Hiranmay Ghosh

Shalini Mukhopadhyay

Venkatramanan Siva Subrahmanian

Sunil Kumar Kopparapu

Indrajit Bhattacharya

Rajan Mindigal Alasingara Bhattachar

Manoj Karunakaran Nambiar

Oishee Mazumder

Shameemraj Mohinuddin Nadaf

Madhurima Ganguly

Srinivasan Jayaraman

Karan Bhavsar

Deb Kumar Ghosh

Subhrojyoti Roy Chaudhuri

Ruddra Dev Roychoudhury

Abhijit Das

Amit Kumar Agrawal

Angshul Majumdar

Garima Mishra

Tanushree Banerjee

Kriti Kumar

Venkata Subramanian Viraraghavan

Puneet Patwari

Soumyadip Maity

Sayan Paul

Arindam Ray

Dipanjan Das

Dhaval Shah

Existing multistatic configurations of Radar systems requires a direct LoS signal and/or time synchronization among the Radar transmitter and the multistatic distributed Radar receivers. The present disclosure provides a phaseless frequency-modulated continuous-wave multistatic Radar (PFMR) imaging that relaxes requirement of the direct LoS signal and only requires a plurality of parameters of a FMCW signal comprising a chirp signal rate, a carrier frequency and, a period of chirp to be known. Further, it also removes condition of the time synchronization among a plurality of FMCW multistatic distributed Radar receivers. However, because of absence of the time synchronization among a plurality of FMCW multistatic distributed Radar receivers, an unknown random phase offset appears after deramping. The present disclosure eliminates the unknown random phase offset, by performing autocorrelation function on a mixed signal, resulting in a phaseless measurement data corresponding to a plurality of FMCW Radar imaging signals.

Method and systems for phaseless frequency-modulated continuous-wave multistatic radar imaging

This disclosure relates generally to method and system for an adaptive filter based learning model for time series sensor signal classification on edge devices. The adaptive filter based learning model for time series sensor signal classification enables automated-computationally lightweight learning (significant reduction in computational resources) and inferring/classification in real-time or near-real-time on CPU/memory/battery life constrained edge devices. The disclosed techniques for time series sensor signal classification on edge devices characterizes the intrinsic signal processing properties of the input time series sensor signals using linear adaptive filtering and derivative spectrum to efficiently construct the adaptive filter based learning model based on standard classification algorithms for time series sensor signal classification.

Adaptive filter based learning model for time series sensor signal classification on edge devices

This disclosure relates generally to method and system for predicting distance of gazed objects using IR camera. Eye tracking technology is widely used to study human behavior and patterns in eye movements. Existing gaze trackers focus on predicting gaze point and hardly analyzes distance of the gazed object from the gazer or directly classify region of focus. The method of the present disclosure predicts gazed objects distance using a pair of IR cameras placed on either side of a smart glass. The gaze predictor ML model predicts distance at least one gazed object positioned from eye of each subject during systematic execution of a set of tasks. From each pupillary information of each pupil a set of features are extracted which are utilized to classify the gazed object of the subject based on the distance into at least one of a near class, an intermediate class, and a far class.

Method and system for predicting distance of gazed objects using infrared (IR) camera

State of the art mechanisms being used for achieving diagnostic-quality images under low-dose settings for general CT imaging have the disadvantages that CT images are fixed during the optimization process to generate perfusion maps, which can lead to suboptimal CT images with respect to the perfusion maps generated, although they might appear spatially smooth or denoised. The disclosure herein generally relates to Computer Tomography (CT) scanning, and, more particularly, to a method and system for CT image reconstruction. The system performs modelling an optimization problem for joint estimation of a set of structural CT images and a perfusion map, and further solves the optimization problem for the reconstruction of the CT images of a subject.

Methods and system for low dose x ray computed tomography perfusion imaging

Small and compact Deep Learning models are required for embedded AI in several domains. In many industrial use-cases, there are requirements to transform already trained models to ensemble embedded systems or re-train those for a given deployment scenario, with limited data for transfer learning. Moreover, the hardware platforms used in embedded application include FPGAs, AI hardware accelerators, System-on-Chips and on-premises computing elements (Fog/Network Edge). These are interconnected through heterogenous bus/network with different capacities. Method of the present disclosure finds how to automatically partition a given DNN into ensemble devices, considering the effect of accuracy—latency power—tradeoff, due to intermediate compression and effect of quantization due to conversion to AI accelerator SDKs. Method of the present disclosure is an iterative approach to obtain a set of partitions by repeatedly refining the partitions and generating a cascaded model for inference and training on ensemble hardware.

Combining compression, partitioning and quantization of DL models for fitment in hardware processors

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. Thus, considering the optical properties along with the polarization properties enables enhanced, accurate characterization of the transparent object.

Polarimetric characterization of transparent objects

This disclosure relates generally to a method and system for classification of cognitive load (CL) using data obtained from wearable sensors. The disclosed method uses a multi-modal based approach using wrist-worn sensors for real time monitoring of CL in real world scenarios and improves the accuracy of detection of CL. A set of distinguishing features are selected from physiological signals received from the wrist-worn sensors. These features are used for training a classification model for classifying the CL of a patient into a no load or a high load. The set of distinguishing features are selected from domain specific features and signal property based generic features of the physiological signals. The disclosed method is used for classification of CL in scenarios such as to check how the cognitive load of a candidate varies during interviews, to assess the participants workload during online meetings and so on.

Method and system for classification of cognitive load using data obtained from wearable sensors

The disclosure generally relates to scene graph generation. Scene graph captures rich semantic information of an image by representing objects and their relationships as nodes and edges of a graph and has several applications including image retrieval, action recognition, visual question answering, autonomous driving, robotics. However, to leverage scene graphs, computationally efficient scene graph generation methods are required, which is very challenging to generate due presence of a quadratic number of potential edges and computationally intensive/non-scalable techniques for detecting the relationship between each object pair using the traditional approach. The disclosure proposes a combination of edge proposal neural network and the Graph neural network with spatial message passing (GNN-SMP) along with several techniques including a feature extraction technique, object detection technique, un-labelled graph generation technique and a scene graph generation technique to generate scene graphs.

Method and system for scene graph generation

This disclosure relates generally to identification and mitigation of bias while training deep learning models. Conventional methods do not provide effective methods for bias identification, and they require pre-defined concepts and rules for bias mitigation. The embodiments of the present disclosure train an auto-encoder to produce a generalized representation of an input image by decomposing into a set of latent embedding. The set of latent embedding are used to learn the shape and color concepts of the input image. The feature specialization is done by training an auto-encoder to reconstruct the input image using the shape embedding modulated by color embedding. To identify the bias, permutation invariant neural network is trained for classification task and attribution scores corresponding to each concept embedding are computed. The method also performs de-biasing the classifier by training it with a set of counterfactual images generated by modifying the latent embedding learned by the auto-encoder.

Method and system for identifying and mitigating bias while training deep learning models

The present invention relates to a method and system for early detection of COVID-19. Existing methods require data from multiple sensors for training a prediction model whose output is considered as final prediction which is actually the prediction for a particular day or time instance. However, this prediction doesn't detect actual infection of COVID-19 since it requires monitoring the change in health of the user over consecutive days. Embodiments of present disclosure overcome these challenges by a prediction model for COVID-19 which requires only data from Photoplethysmography (PPG) sensor seamlessly collected from a wearable device still able to provide accurate COVID-19 prediction with application of a post processing technique on the predictions of the prediction model. Since COVID-19 symptoms have an effect on heartrate and oxygen saturation which are effectively captured by PPG sensor data, studying these dynamics during infection period gives insights to perform early detection of COVID-19.

Method and system for early detection of COVID-19

This disclosure relates generally to a system and a method for mitigating generalization loss in deep neural network for time series classification. In an embodiment, the disclosed method includes compute an entropy of a timeseries training dataset, and a mean and a variance of the entropy and a regularization factor is computed. A plurality of iterations are performed to dynamically adjust the learning rate of the deep Neural Network (DNN) using a Mod-Adam optimization, and obtain a network parameter, and based on the network parameter, the regularization factor is updated to obtain an updated regularized factor. The learning rate is adjusted in the plurality of iterations by repeatedly updating the network parameter based on a variation of a generalization loss during the plurality of iterations. The updated regularized factor of the current iteration is used for adjusting the learning rate in a subsequent iteration of the plurality of iterations.

System and method for mitigating generalization loss in deep neural network for time series classification

This disclosure relates generally to action recognition and more particularly to system and method for real-time radar-based action recognition. The classical machine learning techniques used for learning and inferring human actions from radar images are compute intensive, and require volumes of training data, making them unsuitable for deployment on network edge. The disclosed system utilizes neuromorphic computing and Spiking Neural Networks (SNN) to learn human actions from radar data captured by radar sensor(s). In an embodiment, the disclosed system includes a SNN model having a data pre-processing layer, Convolutional SNN layers and a Classifier layer. The preprocessing layer receives radar data including doppler frequencies reflected from the target and determines a binarized matrix. The CSNN layers extracts features (spatial and temporal) associated with the target's actions based on the binarized matrix. The classifier layer identifies a type of the action performed by the target based on the features.

System and method for real-time radar-based action recognition using spiking neural network(SNN)

This disclosure relates generally to multi-port multi-functional meta-surface coplanar antenna system. Conventional electronic or mechanical solutions for beam steering incur high installation costs with less performance speed and bulk structures. The present disclosure provides multi-port multi-functional meta-surface coplanar antenna system for beam steering control. The disclosed antenna system enables radiator to have a performance diversity application through beam steering functionalities. The disclosed antenna system provides a minimal design complexity and minimal usage of active or passive lumped components. The disclosed system comprises Gradient Refractive Index Meta-surface (GRIM) and the antenna disposed on the same side of a substrate. Beam steering control is performed using port excitations and controlling the phase between the concerned ports externally.

Multi-port multi-functional meta-surface coplanar antenna system for beam steering control

Gesture recognition is a key requirement for Human Computer Interaction (HCl) 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.

System and method for acoustic based gesture tracking and recognition using spiking neural network

State of art techniques rely of FPGA based approaches when power efficiency is of concern. However, compared to SNN on Neuromorphic hardware, ANN on FPGA requires higher power and longer design cycles to deploy neural network on hardware accelerators. Embodiments of the present disclosure provide a method and system for energy efficient hierarchical multi-stage SNN architecture for classification and segmentation of high-resolution images. Patch-to-patch-class classification approach is used, where the image is divided into smaller patches, and classified at first stage into multiple labels based on percentage coverage of a parameter of interest, for example, cloud coverage in satellite images. The image portion corresponding to the partially covered patches is divided into further smaller size patches, classified by a binary classifier at second level of classification. Labels across multiple SNN classifier levels are aggregated to identify segmentation map of the input image in accordance with the coverage parameter of interest.

Energy efficient hierarchical SNN architecture for classification and segmentation of high-resolution images

Growing community of inventors

Kolkata, India

Arpan Pal