Lekkala K.K.,Indian Institute of Information Technology |
Mittal V.K.,Indian Institute of Information Technology
IEEE Region 10 Annual International Conference, Proceedings/TENCON | Year: 2017
In this paper, we propose a robotic video surveillance system. A prototype system is implemented on an aerial vehicle, a quadcopter. It is capable of following a person or any moving object, while simultaneously localizing it, i.e., measuring the coordinates of the quadcopter on a scaled map. The system can operate well, even if it is flying in an unknown and GPS-silent environment. The prototype system consists of 3 major modules. First is an 'Image-based Visual Servoing (IBVS)' module for tracking the desired object or person, and sending the error parameters to the actuators. The second module is 'Parallel Tracking and Mapping (PTAM)' module. The third module is an 'Extended Kalman Filter (EKF)' module. The PTAM and EKF modules help in estimating the location of the quadcopter at every time-instant. The proposed system enables the quadcopter to follow the desired object smoothly, while keeping the tracked object in the center of the frame and maintaining a constant distance from the object. The effectiveness of the system is validated by computing the precise measurements of the aerial movements, while ensuring the minimal error. The proposed method can be helpful for quadcopters used in diverse applications such as surveillance, monitoring and tracking a desired object. © 2016 IEEE.
Verma V.S.,Indian Institute of Information Technology |
Kumar Jha R.,Indian Institute of Technology Patna
Signal, Image and Video Processing | Year: 2015
In this paper, a blind watermarking scheme based on significant difference of lifting wavelet transform coefficients has been proposed. The difference between two maximum coefficients in a block is called as significant difference. Embedding of binary watermark has been done based on the largest coefficient of randomly shuffled blocks of CH3 sub-band. This sub-band is quantized using the predefined threshold value by comparing the significant difference value with the average of significant difference value of all blocks. The watermarked image shows no perceptual degradation as the PSNR value exceeds 42 dB. An adaptive-thresholding-based method is used for watermark extraction. In the proposed technique, the benefit of using lifting wavelet over traditional wavelet is the maximum energy compaction property, which helps in resisting different attacks. The simulation results show higher performance of the proposed technique as compared to the similar existing techniques under different geometric and nongeometric attacks such as amplification, median filtering, sharpening, scaling, rotation, Gaussian noise, salt and paper noise, Gaussian filter and JPEG compression. © 2014, Springer-Verlag London.
Dutt M.,Indian Institute of Information Technology |
Biswas A.,Indian Institute of Science |
Bhattacharya B.B.,Indian Statistical Institute
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) | Year: 2015
The computation of a shortest is othetic path (SIP) between two points in an object is important in various applications such as robotnavigation and VLSI design. However, a SIP between two grid points in a digital object laid on a uniform 2D is othetic square lattice may notbe unique. We assume that each discrete path consists of a sequence ofconsecutive grid edges that starts from the source point and ends at the sink point. In this paper, we present a novel algorithm to calculate the number of such distinct shortest is othetic paths between two given grid points inside a digital object, with time complexity O(S/g2), where S is the total number of pixels in the digital object, and g is the grid size. The number of available SIPs also serves as a metric for shape registration. © Springer International Publishing Switzerland 2015.
Divya D.,Indian Institute of Information Technology |
Agarwal S.,Indian Institute of Information Technology
International Conference on Recent Trends in Information Technology, ICRTIT 2011 | Year: 2011
With ever increasing growth in science and technology, quality of life is improving day by day and health becomes a major concern for everyone. Since many people living in nuclear families, special measures need to be taken especially for the old age people living all alone. Wireless sensor network has become popular due to advancement in the communication capabilities of data, computational processing power and low power microelectronic devices and micro sensors. Body Area sensor network is a collection of wearable sensor nodes that implanted on human body for providing continuous monitoring of health conditions. Basically, we are focusing on how to predict the activity based on sensor reading. This research paper predicts the activity that is based on sensor reading so that we can provide continuous monitoring to patients in order to provide pervasive healthcare. It is in contrast with the traditional event driven approach where a patient visits the doctor only when he is sick. Prediction of activity is done using Support vector Regression (SVR). This paper presents a model to overcome the over-fitting which is due to noise and outliers in dataset. So, we propose an approach which uses weight factor to reduce the prediction error and result in higher accuracy than simple support vector regression. © 2011 IEEE.
Sadagopan S.,Indian Institute of Information Technology
IT Professional | Year: 2012
IT in India has seen amazing growth, thanks to the liberalization of economic policies, conducive government policies, growth in higher education, and the uptake in entrepreneurship. However, unless Indian education and research (on the supply side), IT companies (on the demand side), and the government and media (enablers) continue to innovate, India might not be able to sustain its current leadership position. This article is part of a special issue on IT in Emerging Markets. © 1999-2012 IEEE.
Jain P.,Indian Institute of Information Technology
ACM International Conference Proceeding Series | Year: 2015
A sensor network deployed at international borders requires extreme level of security. The sensor network is supposed to be forensic ready so that at the time of crime and after the crime proper proofs can be displayed before the court of law. The current research work deals with an intelligent multi agent system deployed at multiple sinks in a wireless sensor network. The sensed data is made forensic ready so that it can presented before the court of law. © 2015 ACM.
Mangalraj P.,Indian Institute of Information Technology |
Agrawal A.,Indian Institute of Information Technology
Procedia Computer Science | Year: 2015
A systematic approach for supervised classification of remote sensing images is introduced in this letter. The proposed method deals with the Multi-Level Manifolds, which primarily deals by preserving the local information inside a class along with the class label information. The sharing features are also considered while training the data to represent the parent manifold. The out of sample problem is solved by using Pulse Coded Neural Network which potentially reduces the computational cost. The proposed method solves the major problems of supervised learning systems such as out of sample and preserving local structure. The proposed system is tested in the standard data sets and the results are appreciable. © 2015 The Authors.
Kumar P.,Indian Institute of Information Technology |
Agrawal A.,Indian Institute of Information Technology
Procedia Computer Science | Year: 2015
The advances in Computed Tomography (CT) and Magnetic Resonance Imaging (MRI) scanning techniques are improving the resolution and size of the volume datasets. The prevalence of three dimensional volumetric data is rapidly expanding with the internal information in the different resolution sizes of the dataset. In this paper we proposed an approach that can visualize the inner organs structure of the visible human male dataset in Multi-coordinate Viewing (MCV) framework. This can help to medical experts that are able to peer inside anatomy of the human medical dataset. The volume rendering part has been carried out by the utilization of enhanced ray casting algorithm for the crossing points of 3D square strategy for voxels. We present this system using Graphics Processing Unit or GPU-accelerated Compute Unified Device Architecture (CUDA) based approach for the focusing a specific region while zooming operation. The final results would allow the doctors to diagnose and analyze the atlas of 8-bit CT-scan data using three dimensional visualization with the efficient frame rate rendering speed in multi-operations like zooming, rotating, dragging. The framework is tested for visible human male dataset prepared by National Library of Medicine (NLM, USA) of size 1.2 GB. © 2015 The Authors.
Selvajyothi K.,Indian Institute of Information Technology |
Janakiraman P.A.,Indian Institute of Technology Madras
IEEE Transactions on Power Delivery | Year: 2010
A simple and robust procedure with good transient response is proposed for d-q voltage control in conjunction with harmonics reduction in the output of stand-alone single-phase inverters. The distortion in the load voltage waveform is mainly due to the voltage drop in the filter-inductor caused by the harmonic components of the current. For extracting the in-phase and quadrature signals from harmonic-rich periodic waveforms, a composite observer has been developed, which is faster and more accurate than a simple observer. Separate observers have been provided for the voltage and current signals. Feed forward compensation has been obtained by using the quadrature components derived from the composite current observer. Further, an Inverter could be modeled as a feed back control system with the fundamental component as the desired output and the harmonics as the noise creeping into the output. The well-known control strategy of using a large feed back around the noise signal can be employed to reduce its effect at the output, exhibiting low total harmonic distortion under non-linear loads. The controllers can be easily modeled in the discrete-time domain and the ideas can be directly implemented digitally for single-phase inverters controlled by FPGA or DSP chips. © 2010 IEEE.
News Article | December 5, 2016
SHERMAN OAKS, Calif.--(BUSINESS WIRE)--LendingUSA, a leading provider of point of sale financing solutions, has announced Manoj Mathew as its new Chief Technology Officer. Manoj, a widely recognized and influential member of the marketplace lending technology community, is a seasoned leader with over 20 years of experience working with IT and product engineering companies. An alumnus of the prestigious Indian Institute of Information Technology, Manoj specializes in the execution of strategic initiatives that align customer and future market demands, and has obtained 7 U.S. Patents, with several more patent pending applications. As CTO, Manoj will apply his combined engineering and business background to provide thought leadership and guidance for LendingUSA, which uses a proprietary lending solution that allows customers to apply for financing directly at the point of sale and receive a decision within seconds. “Manoj is a brilliant technologist and fintech visionary with a deep understanding and passion for next-generation technology, problem solving, and market dynamics,” says Camilo Concha, CEO of LendingUSA. “He has invaluable perspective and insight on how to use technology to power progress, improve the user experience, and seamlessly meet the customer’s needs. We’re thrilled to be working with him.” His previous experience includes being the co-founder and Chief Technology Officer of LendFoundry, a fintech accelerator platform for marketplace lending. He was also a co-founder and senior executive at Sigma Infosolutions, a leader in business and IT solutions for both large and small enterprises. During his time at Sigma, he was instrumental in the design and architecture of the origination platform for loanDepot, QuickBridge and the servicing system deployed for Rocketloans. Additionally, as a Senior Systems Engineer, Manoj was a development lead at Finacle, a core banking product by Infosys, which serves 450 million customers across 84 countries. He has also been involved with several computer vision, GPGPU-based image processing and machine learning initiatives at Canon Information and Imaging Solutions. He writes regularly about marketplace lending and has spoken about fintech at major industry conferences including Lendit. “LendingUSA is one of the key innovators in the marketplace lending community, and being a part of a dynamic team with a transformational vision challenges the norm,” says Manoj of the appointment. “They’re using technology in a way that revolutionizes financing for both the merchant and the borrower, and I’m excited to work with them on optimization and strategy as they continue to expand their footprint in the point of sale financing industry.” LendingUSA specializes in point of sale and direct-to-consumer financing solutions with a special focus on a wide range of medical services and consumer goods. The company’s proprietary technology combines traditional and next generation underwriting and fraud detection techniques to approve loan applications in real-time – often in a matter of seconds. LendingUSA builds strong relationships with its merchant partners by providing the tools and training necessary for them to provide the appropriate financing to customers at their point of need. All loans are made by Cross River Bank, a New Jersey State Chartered bank, Member FDIC.