531 research outputs found
Landmarks for clothing retrieval
Get PDFClothing Retrieval is a task that is increasingly becoming popular with the rise of online shopping and social media’s popularity. We propose to solve the clothing retrieval problem using landmarks based on the clothing type and features surrounding the landmarks to get a more ingrained view of the design. We compare this method with other models most of which use the whole image as inputs and show the superiority of the model which gives importance to the crucial parts of the images. For the blouses sub-set from of the Deep Fashion dataset[1], we get an 16% increase in the accuracy for the top 3, 14% in top 5 and 11% top 10 retrieval results using the keypoints extraction methods combined with whole images compared to whole images as inputs. We also observe that the clothes retrieved are more similar in terms or design as well as high level properties like sleeve sizes, folded v/s non-folded sleeves etc.Submission original under an indefinite embargo labeled 'Open Access'. The submission was exported from vireo on 2019-08-22 without embargo termsThe student, Shubham Jain, accepted the attached license on 2019-04-19 at 19:50.The student, Shubham Jain, submitted this Thesis for approval on 2019-04-19 at 19:54.This Thesis was approved for publication on 2019-04-22 at 12:40.DSpace SAF Submission Ingestion Package generated from Vireo submission #13803 on 2019-08-22 at 14:45:13Made available in DSpace on 2019-08-23T20:00:09Z (GMT). No. of bitstreams: 2
JAIN-THESIS-2019.pdf: 36236076 bytes, checksum: ce3af32b75b712d12ebf83d4fb31354f (MD5)
LICENSE.txt: 4209 bytes, checksum: 065a7a1e099c34c652d4dfcf2dcc1ec4 (MD5)
Previous issue date: 2019-04-2
Replication Data for: A HAWQS-SWAT-LSTM Framework for Hydrologic Predictions in Ungauged Watersheds
No full textData for training LSTM models for daily streamflow prediction developed using HAWQS for 531 CAMELS basins in the United States
GIS data for TXSELECT Version 1.0
No full textThis repository serves as a comprehensive data archive for GIS data utilized in the development of TXSELECT (tx.select.tamu.edu). Contents include raw, processed, and intermediate GIS datasets (watershed boundaries, land cover, soil type, census blocks etc.), used to create input files for TXSELECT using the code available at this site - https://github.com/shubhamjain15/TX-SELECT
Replication Data for: Enhancing Prediction and Inference of Daily In-stream Nutrient and Sediment Concentrations using an Extreme Gradient Boosting based Water Quality Estimation Tool - XGBest
No full textThis repository serves as a comprehensive data archive for our paper - "Enhancing Prediction and Inference of Daily In-stream Nutrient and Sediment Concentrations using an Extreme Gradient Boosting based Water Quality Estimation Tool - XGBest". The associated code is available at - https://github.com/GEM-TAMU/XGB-WQ-Predictio
Is Heisenberg's uncertainty principle universal?
No full textHeisenberg’s uncertainty principle is a founding pillar of quantum mechanics. This paper questions the foundations of the quantum mechanics with a thought experiment. The author has come up with this experiment and have found out that there are some anomalies in quantum measurement. Contact Linkedin: https://www.linkedin.com/in/shubham-ambokar-10528b170/To reach out to me: https://www.linkedin.com/in/shubham-ambokar-10528b170
Effect of struts and central tower on aerodynamics and aeroacoustics of vertical axis wind turbines using mid-fidelity and high-fidelity methods
Get PDFThis study investigates the impact of struts and a central tower on the aerodynamics and aeroacoustics of Darrieus Vertical Axis Wind Turbines (VAWTs) at chord-based Reynolds numbers of 8.12 × 104. A 2-bladed H-Darrieus VAWT is used, featuring a 1.5m diameter, a solidity of 0.1 and a blade cross-section of symmetrical NACA 0021. The turbine design is kept simple and straight-bladed which is essential for isolating and analyzing the specific effects of struts and a tower. The high-fidelity Lattice Boltzmann Method (LBM) in PowerFLOW 6-2020 and the mid-fidelity Lifting Line Free Vortex Wake (LLFVW) method in QBlade 2.0 are employed, with the mid-fidelity method providing a faster analytical tool for insights into the turbine performance. Firstly, both the LLFVW (mid-fidelity) and LBM (high-fidelity) methods effectively capture the general trends observed in VAWT power performance. However, the former predicts mean thrust values that are approximately 10% higher, and mean torque values that are approximately 19% higher, in comparison to the latter. Subsequently, the former predicts lower streamwise wake velocities relative to those predicted by the latter. These differences increase in configurations that include struts and a tower (to 30% - 31%). Secondly, the presence of struts and a tower leads to a reduction in both mean power (by 15% to 55%) and thrust (by 3% to 3.6%), with a further small decrease observed when doubling the tower diameter (power and thrust both by 0.5% to 3%). The struts predominantly affect the spanwise distribution of blade loading, while the tower impacts the azimuthal variation of blade loading. Additionally, the addition of struts and a tower reduces low-frequency noise (50-200 Hz) while increasing high-frequency noise (> 300 Hz). The observed decrease in mean blade loading results in reduced low-frequency noise, while the increase in high-frequency noise is ascribed to the increased intensity of BWI/BVI leading to higher unsteady loading fluctuations on blades.Wind Energ
Computational aeroacoustic study of co-rotating rotors in hover
Get PDFThis paper aims to investigate, by means of Lattice-Boltzmann simulations,
the flow-field and far-field noise of two co-axial co-rotating rotors operating
at 3000 rpm in hover conditions. The two co-rotating configurations are
made by 2 × 2-bladed rotors with a fixed axial separation and two different
azimuthal separations ∆φ equal to 84◦ and 12◦
. Isolated 2- and 4-bladed
rotors, are also simulated at the same operating conditions and used as aerodynamic and aeroacoustic reference. For both ∆φ = 84◦ and 12◦
, the upper
rotor tip vortices are accelerated downstream due to the induction from the
lower rotor, avoiding blade vortex interaction (BVI). The lower rotor tip vortices convect into the wake with a lower velocity, causing BVI for ∆φ = 12◦
.
The lower rotor shows a reduction of thrust, relative to the upper rotor, of
36% and 66% for ∆φ = 84◦ and 12◦
, respectively. For ∆φ = 12◦
, the lower
blades act as a wing flap for the upper ones, increasing their thrust. The
∗Corresponding author: PhD Candidate, Flow Physics and Technology Department,
Delft University of Technology
E-mail address: [email protected]
Preprint submitted to Aerospace Science and Technology (AESCTE) July 10, 2024
tonal noise emission for the co-rotating rotors is driven by the interference
between the acoustic waves from upper and lower rotors. Because of destructive interference, the configuration ∆φ = 84◦
shows a first harmonic up to 15
dB lower than ∆φ = 12◦
, but still 4.5 dB higher than the isolated 4-bladed
roto
Using Optical See-through Mixed Reality for Enhanced Shopping Experience in Omnichannel Retail
No full textThis cumulative doctoral thesis is focused on the digital technology: ‘optical see-through mixed reality’ and its applications in omnichannel retail. The research aims at designing a mixed reality-based shopping assistant application that can be deployed in omnichannel retail environments to help retail businesses in the current age of digital retail. The retail sector has been transitioning into a newer model i.e., omnichannel retail where providing customers with an enhanced shopping experience is one of the core ideas of the concept. As this trend grew, retailers have started to implement different digital technologies in their ecosystems to achieve this enhanced shopping experience. Among these technologies, immersive environments such as optical see-through mixed reality have emerged as a potential tool that can elevate shoppers’ experiences during their shopping journey as a result of the technology’s various capabilities that include combining physical and digital environments. Now, the current literature has pointed out a lot of advantages of the technology including advantages specifically in retail environments but lacks knowledge around how this technology can be deployed in omnichannel retail environments. As the omnichannel retail ecosystem is customer-centric, this knowledge should arise from customer perceptions towards mixed reality artifacts in retail environments.
Towards this, the overall research is aimed at designing a customer-centric optical see-through mixed reality-based shopping assistant system application that has been developed and optimized for mixed reality headsets: Microsoft HoloLens and HoloLens 2 as the archetype. The research uses the design science framework to design, develop and test the physical artifact as a proof of concept. The dissertation includes seven different articles that were developed during the research period that aim to help retail managers in understanding the dynamics of the deployment of mixed reality technology in retail. They also aim to help designers and developers in understanding the process of designing mixed reality applications. Finally, it can potentially help researchers to further explore the topic of immersive environments in the domain of human-computer interaction, information systems, customer experience and digital retail.Author Shubham JainDissertation Universität Linz 202
Exploring susceptibility to use demand responsive transport (DRT)
Get PDF© 2016 Shubham JainShared transportation providing point-to-point services on demand, although not an unknown element in urban mobility, has started gaining more presence with growth of information technology in the transport sector. These forms of transport modes will supplement or compete with existing public and private transport. Their mixed reception in the past is a matter of concern especially before making investment decisions. To find feasible opportunities of implementation, an estimation of the demand patterns in the target city is desirable.
This research provides and evaluates a methodology for this estimation that avoids ambivalent and expensive user surveys. Demand patterns are caused by the spatial variation of socio-economic and demographic characteristics, family structures, and travel behavior over the city. Thus, the new methodology takes into account the use of socio-economic and demographic data and current trip data from travel surveys of a sample of the population, along with usage patterns of existing similar services elsewhere in the world. Demographic factors such as gender, age, occupation, income, household structure, motor vehicle ownership, and driving licence availability together with trip characteristics such as current trip purpose, walking time, and waiting time can be analyzed to come up with demand patterns, and their variation in the city. Usage patterns from existing similar services worldwide are then used to explore the overall spatial pattern of susceptibility of DRT in a target city. The outcomes identify more favorable areas for implementation of DRT.
The methodology can be validated by applying it on existing transport modes in the target city which will also help in understanding the nature of competition among the proposed and existing transport modes. As the review of operating services is generic, it can be used in conjunction with respective travel surveys in different places. Similarly, review can be done for any proposed transport mode, and provided methodology can be applied for exploring demand patterns. The methodology is tested for Greater Melbourne in this study.
Further, synthetic population is created at household and person levels for Greater Melbourne. PopSynWin, which is based on Iterative Proportional Fitting (IPF) algorithm, and PopGen, which is based on Iterative Proportional Update (IPU) algorithm, are used
as tools for this purpose, generating two different synthetic populations. Both the generated populations are compared statistically, and the better one is used to assign travel diaries from sample travel surveys for a study region. The methodology to explore the susceptibility to use DRT, provided in this research, is applied on individual travel diaries to further explore demand patterns at a finer granularity
Design of inertial and camera sensing support for smart intersections
No full textModern cities are alive with sensors, including but not limited to smartphones, cameras, vehicles, and wearable devices. Contrary to popular belief that the evolution of smart cities needs an overhaul of advanced sensors across our cities, this dissertation presents techniques that enable existing sensing devices to expand their role and innovate novel smart city context. We undertake the task of supporting a diverse set of applications, ranging from large-scale video analytics to pedestrian safety, on a heterogeneous assembly of sensors. Motivated by rising pedestrian fatalities in our cities, we investigate the performance of GPS-based approaches for determining pedestrian risk in dense urban environments. To address its inadequacy, we introduce a novel outdoor surface profiling technique using shoe-mounted inertial sensors for location classification based on surface gradient profile and step patterns. We seek to detect transitions from sidewalk locations to in-street locations, to enable applications such as alerting texting pedestrians when they step into the street. We achieve transition detection rates higher than 95% even in the intricate midtown Manhattan pedestrian environment. Further, we extend this ability to mobile cameras, and explore how well commercial-off-the-shelf smartphone cameras can learn texture to distinguish among paving materials in uncontrolled outdoor urban settings. We devise an approach that performs material recognition on the pedestrian's walking surface, with more than 85% accuracy, to identify safe and unsafe walking locations. Finally, to advance the state of video analytics in smart cities, we build a virtualization system for public cameras to support multiple applications simultaneously. We introduce the concept of mobility-awareness, which enables these otherwise static cameras to pan, tilt, and zoom to capture events of interest. This improves immensely upon the current state-of-the-art wherein traffic operators examine live video streams. Experiments with a live camera setup demonstrate that we can support multiple applications simultaneously, capturing up to 80% more events of interest in a wide scene, compared to a fixed view camera. This work is based on the insight that relative positions and motion patterns are crucial for generating safety context and meaningful analytics at traffic intersections. Furthermore, we demonstrate the efficacy of our approaches by building end-to-end systems, calling for exhaustive real-world data collection in complex metropolitan environments like New York, London, and Paris; supported by rigorous testing and scalability of the solution.Ph.D.Includes bibliographical referencesby Shubham Jai
- …
