Nitte Meenakshi Institute of Technology

About: Nitte Meenakshi Institute of Technology is a based out in . It is known for research contribution in the topics: Computer science & Ultimate tensile strength. The organization has 846 authors who have published 644 publications receiving 2702 citations.

Multimodal image feature detection with ROI-based optimization for image registration

Abstract: Image registration plays an imperative part of multimodal video analysis system. In video surveillance applications, change in the environmental conditions makes the registration process hard. Use of multiple sensors makes the system more robust to environmental changes as compared to single sensor imaging system. Using multiple modalities such as infrared(IR)/thermal sensors and CMOS image sensors augment the sturdiness of the surveillance system. Here we propose hardware implementation of feature detection on Genesys 2 Kintex-7 FPGA for a multimodal surveillance system, which is robust in poor lighting conditions and affine changes. To reduce the processing time, a region of interest (ROI) is identified and feature extraction is performed in this region. Design optimization in hardware architecture resulted in achieving the real-time performance of image registration on HD 720p video.

Habitability classification of exoplanets: a machine learning insight

Abstract: We explore the efficacy of machine learning (ML) in characterizing exoplanets into different classes. The source of the data used in this work is University of Puerto Rico’s Planetary Habitability Laboratory’s Exoplanets Catalog (PHL-EC). We perform a detailed analysis of the structure of the data and propose methods that can be used to effectively categorize new exoplanet samples. Our contributions are twofold. We elaborate on the results obtained by using ML algorithms by stating the accuracy of each method used and propose a paradigm to automate the task of exoplanet classification for relevant outcomes. In particular, we focus on the results obtained by novel neural network architectures for the classification task, as they have performed very well despite complexities that are inherent to this problem. The exploration led to the development of new methods fundamental and relevant to the context of the problem and beyond. The data exploration and experimentation also result in the development of a general data methodology and a set of best practices which can be used for exploratory data analysis experiments.

PCVOS: Principal component variances based off-line signature verification

Abstract: Offline signature verification system is widely used as a behavioral biometric for identifying a person. This behavioral biometric trait is a challenge in designing the system that has to counter intrapersonal and interpersonal variations. In this paper, we propose a novel technique PCVOS: Principal Component Variances based Off-line Signature Verification on two critical parameters viz., the Pixel Density (PD) and the Centre of Gravity (CoG) distance. It consists of two parallel processes, namely Signature training which involves extraction of features from the samples of database and Test signature analysis which performs extraction of features from the test samples. The trained values from the database are compared with the features of the test signature using Principal Component Analysis (PCA). The PCVOS algorithm shows a notable improvement over the algorithms in [21], [22] and [23].