Faculty Publications
Permanent URI for this communityhttps://idr.nitk.ac.in/handle/123456789/18736
Publications by NITK Faculty
Browse
14 results
Search Results
Item A multi-space approach to zero-shot object detection(Institute of Electrical and Electronics Engineers Inc., 2020) Gupta, D.; Anantharaman, A.; Mamgain, N.; Kamath S․, S.; Balasubramanian, V.N.; Jawahar, C.V.Object detection has been at the forefront for higher level vision tasks such as scene understanding and contextual reasoning. Therefore, solving object detection for a large number of visual categories is paramount. Zero-Shot Object Detection (ZSD) - where training data is not available for some of the target classes - provides semantic scalability to object detection and reduces dependence on large amount of annotations, thus enabling a large number of applications in real-life scenarios. In this paper, we propose a novel multi-space approach to solve ZSD where we combine predictions obtained in two different search spaces. We learn the projection of visual features of proposals to the semantic embedding space and class labels in the semantic embedding space to visual space. We predict similarity scores in the individual spaces and combine them. We present promising results on two datasets, PASCAL VOC and MS COCO. We further discuss the problem of hubness and show that our approach alleviates hubness with a performance superior to previously proposed methods. © 2020 IEEE.Item Modeling Uber Data for Predicting Features Responsible for Price Fluctuations(Institute of Electrical and Electronics Engineers Inc., 2022) Sindhu, P.; Gupta, D.; Meghana, S.; Anand Kumar, M.In the field of economics, the features and patterns of the transportation system, including classical modes of transportation such as subways and taxis, as well as innovative tools such as car pooling platforms(Uber, Lyft, etc), are key research topics. The study here demonstrates how an Uber dataset is, which comprises Uber's New York City data, works. Uber is an online service provider platform via internet or a mobile application that avails ride-hailing service. In essence, it matches passengers with drivers of vehicles to book a ride from one place to another. The service connects users with drivers who will drive them to their desired location. The dataset contains primary data about Uber pick-ups, including the date, time, longitude, and latitude coordinates. The paper attempts to examine data from different locations, weathers, hours, and dates (intraday and midweek) in New York City and apply time series data analysis, statistical regression on the dataset, and predict Uber ride prices. We arrive at conclusions by analyzing data using various graphs, calculating and estimating the influence of these elements on Uber riders' payment amounts, and emphasizing features that cause price fluctuation. © 2022 IEEE.Item NITK_LEGAL at SemEval-2023 Task 6: A Hierarchical based system for identification of Rhetorical Roles in legal judgements(Association for Computational Linguistics, 2023) Sindhu, P.; Gupta, D.; Meghana, S.; Anand Kumar, M.The ability to automatically recognise the rhetorical roles of sentences in a legal case judgement is a crucial challenge to tackle since it can be useful for a number of activities that come later, such as summarising legal judgements and doing legal searches. The task is exigent since legal case documents typically lack structure, and their rhetorical roles could be subjective. This paper describes SemEval-2023 Task 6: LegalEval: Understanding Legal Texts, Sub-task A: Rhetorical Roles Prediction (RR). We propose a system to automatically generate rhetorical roles of all the sentences in a legal case document using Hierarchical Bi-LSTM CRF model and RoBERTa transformer. We also showcase different techniques used to manipulate dataset to generate a set of varying embeddings and train the Hierarchical Bi-LSTM CRF model to achieve better performance. Among all, model trained with the sent2vec embeddings concatenated with the handcrafted features perform better with the micro f1-score of 0.74 on test data. The dataset utilised in our task is available at 1 © 2023 Association for Computational Linguistics.Item DETECTION OF BUILDING INFRASTRUCTURE CHANGES FROM BI-TEMPORAL REMOTE SENSING IMAGES(Institute of Electrical and Electronics Engineers Inc., 2024) Sravya, N.; Kevala, V.D.; Akshaya, P.; Basavaraju, K.S.; Lal, S.; Gupta, D.Change detection (CD) from satellite images is crucial for Earth observation, especially in monitoring urban growth patterns. Recent research has largely focused on using Deep Learning (DL) techniques, particularly variations of Convolutional Neural Network (CNN) architectures. While DL methods have shown promise, many of the models could not preserve the changed areas shape and it fails in predicting the correct edges of changed areas. This paper introduces a CNN based Building Infrastructure Change Detection Network (BICDNet) for predicting changes from bi-temporal remote sensing images. The model leverages a modified Fully Convolutional Siamese-Difference Network to extract detailed features from given images which includes a Multi-Feature Extraction (MFE) block designed to capture features from changed areas of various size within the given input images. To further refine these feature pairs, a modified Atrous Spatial Pyramid Pooling (MASPP) module is integrated, which effectively captures contextual information at multiple scales. The comparison study shows that the proposed BICDNet performs better than the existing CD models. © 2024 IEEE.Item Sodium ion incorporated alumina - A versatile anisotropic ceramic(Elsevier Ltd, 2019) Pujar, P.; Gupta, B.; Sengupta, P.; Gupta, D.; Mandal, S.The present article is a review of crystal structure dependent anisotropic properties of β and β″-phases of sodium ion incorporated alumina. The anisotropy in electrical properties such as ionic conductivity and dielectric permittivity is due to the layered structure. Conducting plane between two consecutive spinel aluminas constituting loosely bound mobile sodium ions, promote ionic conductivity in the parallel direction. In contrary, the restricted movement of ions in the orthogonal direction brings about polarization giving it directional dielectric property. High ionic conductivity of 1.3 S/cm and large dielectric constant of ˜ 200 are reported. Exchanging sodium ions with different cations, such as potassium and lithium, results in similar anisotropy. The processing of β and β″-phases along with metastability of intermediate mullite phase is described in the current review. In addition, the applications of sodium ion incorporated aluminas, such as solid electrolyte in batteries, thin film transistors and gas sensors are discussed. © 2019 Elsevier LtdItem Performance comparison of MANET routing protocols(2010) Kumar, J.; Kulkarni, M.; Gupta, D.A Mobile Ad hoc Network (MANET) is a collection of wireless mobile nodes forming a temporary network without using any existing infrastructure. Since not many MANETs are currently deployed, research in this area is simulation based. We have compared different routing protocols like DSR, DSDV, AODV to study the mobility characteristics of different scenarios in which MANETs may be deployed. Evaluation of the impact of different mobility models on the performance of these routing protocols is carried out, based upon different parameters-throughput, Packet delivery ratio, end to end delay, and offered traffic load. Also, mobility models are compared based on these performance parameters in various routing protocols. © 2010 Praise Worthy Prize S.r.l. - All rights reserved.Item Development of low temperature stoichiometric solution combustion derived transparent conductive ternary zinc tin co-doped indium oxide electrodes(Royal Society of Chemistry, 2017) Pujar, P.; Gandla, S.; Singh, M.; Gupta, B.; Tarafder, K.; Gupta, D.; Noh, Y.-Y.; Mandal, S.Here, the development of transparent conductive zinc tin co-doped indium oxide (IZTO: In1.4Sn0.3Zn0.3O3) ternary electrodes is addressed through low temperature solution combustion processing. Optimization of fuel to oxidizer ratio offers low temperature (?130 °C) of combustion with balanced redox reaction. The thin films of IZTO annealed at different temperatures showed a decreasing trend in the resistivity with a fixed order of 10-2 ? cm and the film with a highest Hall mobility of 5.92 cm2 V-1 s-1 resulted at 400 °C. All the films with different temperatures of annealing were smooth (rms ? 2.42 nm) in nature and the IZTO film annealed at 200 °C is 83% transparent in the visible spectra. The effective band gap of 0.9 eV determined from first-principles density functional theory gives clear evidence for the conducting nature of IZTO. The thin film transistor fabricated with IZTO as a gate electrode with poly(methyl methacrylate) and pentacene as the dielectric and channel material, respectively, exhibited a saturation mobility of 0.44 cm2 V-1 s-1 and Ion/Ioff ratio of 103. Further, the printability of the IZTO combustible precursor is established which resulted in anti-edge deposition of the printed feature. © 2017 The Royal Society of Chemistry.Item Retention of high dielectric constant sodium beta alumina via solution combustion: Role of aluminum ions complexation with fuel(Elsevier Ltd, 2018) Gupta, B.; Pujar, P.; Mal, S.S.; Gupta, D.; Mandal, S.In the present study, solution combustion technique has been explored to synthesize Sodium ?-alumina (SBA; NaAl11O17) powder and thin films. Three fuels namely urea, glycine and citric acid have been used to seek the feasibility of synthesizing crystalline SBA powder at low temperature. Also, the effect of nature of fuels used as well as calcination treatment on phase evolution and morphology of the as-combusted powder was investigated. Thermal analysis and X-ray diffraction studies suggest the formation of crystalline SBA powder at temperature as low as 259 °C, using urea in the combustion reaction whereas other fuels resulted in amorphous SBA phase and this variation in phase was found due to difference in exothermicity of the fuel used. Thermodynamic and spectroscopic analyses showed that the exothermicity of fuel depends on various factors like (i) standard heat of formation of fuel and (ii) the complexation offered by fuel to metal cations. Furthermore, sodium ?-alumina thin film capacitor (metal-insulator-metal) was also fabricated using urea via spray combustion synthesis. The sodium ?-alumina thin film showed a high dielectric value (?r) of ~21. © 2017 Elsevier Ltd and Techna Group S.r.l.Item Image quality restoration framework for contrast enhancement of satellite remote sensing images(Elsevier B.V., 2018) Suresh, S.; Das, D.; Lal, S.; Gupta, D.Researches in satellite remote sensing images mainly revolves around enhancement of contrast and removal of noise in image, which affects the data comprehensibility and clarity. Hence, it is always a challenge to process the satellite remote sensing images in order to obtain better quality images with enhanced visibility and minimum image artifacts for improving their application value. In this paper, an effective quality enhancement framework is proposed, which mainly focuses on contrast enhancement of satellite remote sensing images. Several satellite remote sensing images were tested to ratify the effectiveness of the proposed method over other existing remote sensing enhancement methods and their quantitative results are borne out by NIQMC (No Reference Image Quality Metric for Contrast distortion), BIQME (Blind Image Quality Measure of Enhanced images), MICHELSON (Michelson Contrast), DE (Discrete Entropy), EME (Measure of enhancement) and PIXDIST (Pixel distance) along with qualitative results comparison. Results depict that the visual enhancement obtained using the proposed method is superior to other existing enhancement methods. Finally, the simulation results unveil that proposed method is effective and efficient for satellite remotes sensing images. © 2018 Elsevier B.V.Item A balancing between super transparency and conductivity of solution combustion derived titanium doped indium oxide: Effect of charge carrier density and mobility(Elsevier B.V., 2018) Pujar, P.; Vardhan, R.V.; Gupta, D.; Mandal, S.In this contribution, super transparent (~100%) and conducting In14Ti1O23 (Titanium doped Indium oxide; InTiO) films were reported via solution combustion processing with acetylacetone as fuel. Both bulk-powder and thin film systems were studied and revealed the efficacy of low temperature combustion synthesis which yielded crystalline InTiO powder at 150 °C and its film counterpart had shown pronounced crystalinity with temperature. Also, all films with varying annealing temperature were smooth with rms value ranging from 0.29 nm to 1.9 nm. In addition, the charge carrier density in all films found to be of the order 1019 cm?3, possessing highest transparency nearly equals to uncoated glass at an annealing temperature of 350 °C having maximum of ~67% metal-oxygen-metal framework (or lattice oxygen) confirmed via X-ray photoelectron spectroscopy. Also, the highest conductivity of 20 S/cm at an annealing temperature of 450 °C clearly conveyed the potential of solution combustion processing in the fabrication of ultra-transparent InTiO films with no sophistication in the film fabrication. © 2018