Open Access. Powered by Scholars. Published by Universities.®
- Discipline
-
- Electrical and Computer Engineering (18)
- Mechanical Engineering (7)
- Power and Energy (7)
- Civil and Environmental Engineering (5)
- Bioresource and Agricultural Engineering (4)
-
- Civil Engineering (3)
- Computer Engineering (3)
- Materials Science and Engineering (3)
- Physical Sciences and Mathematics (3)
- Biomedical Engineering and Bioengineering (2)
- Life Sciences (2)
- Aerospace Engineering (1)
- Agriculture (1)
- Chemical Engineering (1)
- Computer Sciences (1)
- Electrical and Electronics (1)
- Energy Systems (1)
- Environmental Engineering (1)
- Environmental Sciences (1)
- Food Science (1)
- Geography (1)
- Nanoscience and Nanotechnology (1)
- Operations Research, Systems Engineering and Industrial Engineering (1)
- Physics (1)
- Remote Sensing (1)
- Social and Behavioral Sciences (1)
- Structural Engineering (1)
- Structures and Materials (1)
- Transportation Engineering (1)
- Keyword
-
- Electrospinning (2)
- Lithium metal battery (2)
- 3D Current collector (1)
- 3D Printing (1)
- ANSYS (1)
-
- Accelerated Bridge Construction (1)
- Accuracy Ensured Cmpression (1)
- Additive Manufacturing (1)
- Additive material (1)
- Adhesive (1)
- Aerospace (1)
- Agricultural Spraying (1)
- Alternative Energy (1)
- Aluminum (1)
- Artificial Neural Network (1)
- Artificial SEI (1)
- Asphalt (1)
- Atomic Force Microscopy (1)
- Back Pain (1)
- Biomedical Devices (1)
- Bioretention cell/Rain garden (1)
- Bone Tissue Engineering (1)
- Brachiation Robot (1)
- Bridge Columns (1)
- CFD (1)
- CNN Pruning (1)
- Cellulose Acetate (1)
- Cellulose Nanofibers (1)
- Charge carrier dynamics (1)
- Cold Spray (1)
Articles 1 - 30 of 30
Full-Text Articles in Engineering
Strain-Based Mechanical Failure Analysis Of Buried Steel Pipeline Subjected To Landslide Displacement Using Finite Element Method, Sean Goodluck Mamuzo Edeki
Strain-Based Mechanical Failure Analysis Of Buried Steel Pipeline Subjected To Landslide Displacement Using Finite Element Method, Sean Goodluck Mamuzo Edeki
Electronic Theses and Dissertations
Landslide displacement is one of the major threats to the structural integrity of buried oil and natural gas pipelines that are often located far from major markets with terrains prone to permanent ground deformations. These pipelines can experience large longitudinal strains and circumferential deformation resulting from the differential ground movements thereby potentially impacting pipeline safety by adversely affecting structural capacity and leak tight integrity. In order to proffer theoretical basis for the design, safety evaluation and maintenance of pipelines, the failure analysis and mechanical behavior of buried API X65 steel pipeline perpendicularly crossing landslide area was investigated with the Finite …
Two-Dimensional Nanomaterials And Their Composites For Electrochemical Detection Of Toxic Mercury Ions In Water, Md Tawabur Rahman
Two-Dimensional Nanomaterials And Their Composites For Electrochemical Detection Of Toxic Mercury Ions In Water, Md Tawabur Rahman
Electronic Theses and Dissertations
The presence of trace amounts of mercury ion (Hg2+) in drinking water has a detrimental effect on human health. The development of an electrochemical sensor for Hg2+ detection is still challenging to obtain ultra-trace sensitivity, excellent selectivity, wide Linear Detection Ranges (LDRs), and ultra-low detection limit. This work presents an electrochemical sensor based on two-dimensional nanomaterials and their composites for the enhanced sensing of Hg2+ in water. Graphene oxide (GO)-silver nanowires (AgNWs) composite and metallic 1T phase tungsten disulfide (WS2) microflowers were utilized for the fabrication of electrochemical sensors using drop-casting. Under the optimized experimental conditions, …
Grain Boundary Defect Passivation Of Mixed Cation Perovskite Solar Cell To Improve Device Performance, Sheikh Ifatur Rahman
Grain Boundary Defect Passivation Of Mixed Cation Perovskite Solar Cell To Improve Device Performance, Sheikh Ifatur Rahman
Electronic Theses and Dissertations
Rigorous research on perovskite solar cell for more than ten years all over the world has led the rapid development of perovskite technology giving us device efficiency of 25.2%, surpassing the performance of traditional thin-film solar technologies on small scale devices (<1cm2). Perovskite material has excellent electronic and optical properties making it greatly suitable for PV application. The device fabrication process being simple with low cost and solution processability has led the researchers to visualize high efficiency scalable devices. Due to solution process fabrication method the defects in the materials causes the devices perform less than anticipated. These …
Distributed Machine Learning Approach To Fast Frequency Response-Based Inertia Estimation In Low Inertia Grids, Abodh Poudyal
Distributed Machine Learning Approach To Fast Frequency Response-Based Inertia Estimation In Low Inertia Grids, Abodh Poudyal
Electronic Theses and Dissertations
Recent updates to the IEEE 1547-2018 standard allow active participation of distributed energy resources (DERs) in power grid services with the goal of increased grid reliability and resiliency. With the rapid growth of DERs towards a low inertia converter-dominated grid, the DERs can provide fast frequency response (FFR) services that can quickly counteract the change in system frequency through inertial support. However, in low voltage grids, frequency and voltage face dynamics coupling due to a high resistance to reactance ratio and cannot be controlled separately as in the bulk electric grid. Due to the coupling effect, the control of one …
Tailoring The Grain Boundaries Of Wide-Bandgap Perovskite Solar Cells By Molecular Engineering, Khalid Emshadi
Tailoring The Grain Boundaries Of Wide-Bandgap Perovskite Solar Cells By Molecular Engineering, Khalid Emshadi
Electronic Theses and Dissertations
Due to the attraction of fabricating highly efficient tandem solar cells, wide-bandgap perovskite solar cells have attracted substantial interest in recent years. However, polycrystalline perovskite thin-films show the existence of trap states at grain boundaries, which diminish the optoelectronic properties of the perovskite and thus remains a challenge. This research demonstrates a one-step solution-processing of the [MA0.9Cs0.1Pb(I0.6Br0.4)3] wide-bandgap perovskite using Phenylhydrazine Iodide with amino groups to successfully passivate the trap density within grain boundaries and increase the perovskite grain size. The reinforced morphology and grain boundaries treatment considerably enhanced the photovoltaic performance …
Past To Present (P2p): Road Thermal Image Colorization, Yuseong Park
Past To Present (P2p): Road Thermal Image Colorization, Yuseong Park
Electronic Theses and Dissertations
Thermal image colorization into realistic RGB image is a challenging task. Thermal cameras are easily to detect objects in particular situation (e.g. darkness and fog) that the human eyes cannot detect. However, it is difficult to interpret the thermal image with human eyes. Enhancing thermal image colorization is an important task to improve these areas. The results of the existing colorization method still have color ambiguities, distortion, and blurriness problems. This paper focused on thermal image colorization using pix2pix network architecture based on Generative Adversarial Net (GAN). Pix2pix is a model that transforms thermal image into RGB image, but our …
Structural Performance Evaluation Of Dynamic Message Signs With Adhesive And Welded Joints, Ibin Amatya
Structural Performance Evaluation Of Dynamic Message Signs With Adhesive And Welded Joints, Ibin Amatya
Electronic Theses and Dissertations
The ultimate objective of this study was to determine the structural performance of the Dynamic Message Sign (DMS) bonded with chemical adhesive in terms of ultimate strength and fatigue strength. To achieve this objective, this study first investigated the effects of various parameters, including humidity and temperature on tensile, shear, peel, and cleavage strength through small-scale tests of adhesive specimens with variation in specimen width. The tensile, shear, peel, and cleavage specimens were conditioned in a temperature-humidity controlled chamber prior to its testing. The tensile, shear, peel, and cleavage specimens with different widths were tested according to American Society for …
Seismic Performance Of Mechanically Spliced Bridge Columns Through Analytical Studies, Mathew Ray Lavoy
Seismic Performance Of Mechanically Spliced Bridge Columns Through Analytical Studies, Mathew Ray Lavoy
Electronic Theses and Dissertations
Reinforcement continuity in concrete members is traditionally achieved through lap splicing. Mechanical bar splices are an alternative to traditional lap splices. Mechanical bar splices utilize mechanical devices known as bar coupler to connect the ends of reinforcing bars. Bar couplers are common in accelerated bridge construction applications at non-critical members but are not allowed in the plastic hinge region of bridge columns probably due to a lack of knowledge on how they affect seismic performance of columns. A comprehensive analytical study was carried out to determine how bar couplers used at the base of bridge columns affect: 1) column lateral …
Computational Intelligence Approaches For Energy Optimization In Microgrids, Tamal Roy
Computational Intelligence Approaches For Energy Optimization In Microgrids, Tamal Roy
Electronic Theses and Dissertations
The future electrical system termed as smart grid represents a significant paradigm shift for power industry. Nowadays, microgrids are becoming smarter with the integration of renewable energy resources (RESs) , diesel generators , energy storage systems (ESS), and plug-in electric vehicles (PEV or EV) . However, these integration bring with new challenges for intelligent management systems. The classical power generation approaches can no longer be applied to a microgrid with unpredictable renewable energy resources. To relive these problem, a proper power system optimization and a suitable coordination strategy are needed to balance the supply and demand. This thesis presents three …
Different Treatments Of Biochar Carbon Yp-50f To Enhance Specific Capacitance Of Supercapacitor, Ezaldeen Adhamash
Different Treatments Of Biochar Carbon Yp-50f To Enhance Specific Capacitance Of Supercapacitor, Ezaldeen Adhamash
Electronic Theses and Dissertations
High cost and environmentally unfavorable impacts are the significant difficulties that prohibit renewable energy storage from many applications. To work with these problems, novel renewable materials such as biomass derived carbon that have high surface area and low cost should be used. Renewable biochar carbon YP-50, synthesized from coconut, and activated using different plasma gases including methane (CH4), carbon dioxide (CO2), hydrogen (H2), and argon (Ar). Compared with the conventional activation method, the plasma treatment takes less effort and time. Significant improvements of the specific capacitance (SC) and specific surface area (SSA) were observed …
Semantic Segmentation Using Modified U-Net Architecture For Crack Detection, Michael Sun
Semantic Segmentation Using Modified U-Net Architecture For Crack Detection, Michael Sun
Electronic Theses and Dissertations
The visual inspection of a concrete crack is essential to maintaining its good condition during the service life of the bridge. The visual inspection has been done manually by inspectors, but unfortunately, the results are subjective. On the other hand, automated visual inspection approaches are faster and less subjective. Concrete crack is an important deficiency type that is assessed by inspectors. Recently, various Convolutional Neural Networks (CNNs) have become a prominent strategy to spot concrete cracks mechanically. The CNNs outperforms the traditional image processing approaches in accuracy for the high-level recognition task. Of them, U-Net, a CNN based semantic segmentation …
Using Computational Fluid Dynamics To Accurately Model Agricultural Spray Nozzles, Zachary Chapman
Using Computational Fluid Dynamics To Accurately Model Agricultural Spray Nozzles, Zachary Chapman
Electronic Theses and Dissertations
Computational fluid dynamics (CFD) is a tool used by engineers in many industries to study fluid flow. A relatively new industry to adopt the use of CFD is the agricultural industry. The present work seeks to understand whether CFD can be used to accurately model spray nozzles. A spray nozzle commonly used in agricultural spraying was simulated. First, the impact of factors such as mesh size, mesh type, and physics models have on the solution were investigated. Next, a method to pulse the spray was determined. This was required to compare simulation results with experimental data. A user-defined function was …
Understanding The Effect Of Single- And Twin-Screw Extrusion Processing Parameters On Physical And Physico-Chemical Properties Of Sprouted Quinoa And Sprouted Proso Millets, Gabriela John Swamy
Understanding The Effect Of Single- And Twin-Screw Extrusion Processing Parameters On Physical And Physico-Chemical Properties Of Sprouted Quinoa And Sprouted Proso Millets, Gabriela John Swamy
Electronic Theses and Dissertations
According to Mordor Intelligence Research, the new compositional research and moisture content extrusion process are helping the growth of the plant protein market. The demand for plant proteins is growing at a fast rate, owing to change in lifestyle, lack of balanced dietary intake, and improved research and development in order to develop new kinds of plant-protein enriched products. It is necessary to identify the right plant protein sources and choose the right processing methodology to create highly digestible foods which can be consumed by infants and elderly as well. In addition, it is important to support local farmers and …
Development Of Cellulose Acetate-Based Scaffold For Bone Tissue Engineering Applications, Ruhit Sinha
Development Of Cellulose Acetate-Based Scaffold For Bone Tissue Engineering Applications, Ruhit Sinha
Electronic Theses and Dissertations
Cellulose Acetate (CA) based biomaterials are being used as substrates for bone ingrowth applications due to their nontoxic and nonirritant nature coupled with optimum morphology and stiffness. Electrospinning with additives and/or post-treatment has emerged as a viable protocol to further improve mechanical properties of CA and expand its utility. Herein, we highlight the role of potassium chloride (KCl) in association with airdrying to enhance the elastic modulus and tensile strengths of CA fibers. Salt aggregation in between fibers is observed through Scanning Electron Microscopy (SEM), however, Fourier-transform infrared (FTIR) analysis signifies the interactions between K+ ions and acetyl groups. The …
Effects Of Cellulose Nano-Fiber As An Additive On Performance Of Asphalt Binders And Mixes, Marco Paulo Pereira Castro
Effects Of Cellulose Nano-Fiber As An Additive On Performance Of Asphalt Binders And Mixes, Marco Paulo Pereira Castro
Electronic Theses and Dissertations
Findings of a study conducted on asphalt binders and mixes modified by addition of cellulose nanofiber (CNF) to evaluate the feasibility of them as an additive is presented in the current thesis. Cellulose Acetate (CA) is an ester of cellulose and is obtained by the reaction of cellulose with acetic anhydride and acetic acid in the presence of sulfuric acid. Due to its high temperature sensitivity, high ductility, large surface area, high strain resistance, and low electrical resistivity, cellulose acetate can be used for several application. The question is if CNF can be used as an additive in the asphalt, …
Enhanced Charge Carrier Dynamics And Defects Passivation In Hybrid Perovskite Solar Cell Using A Novel Organic Halide Salt, Md Ashuqur Rahman Laskar
Enhanced Charge Carrier Dynamics And Defects Passivation In Hybrid Perovskite Solar Cell Using A Novel Organic Halide Salt, Md Ashuqur Rahman Laskar
Electronic Theses and Dissertations
In the last few years, active research works on hybrid perovskite solar cells (HPSCs) have been going on throughout the world as a part of clean energy revolution. HPSCs have already proved their potentiality as future solar technology through their promising photovoltaic performances and ability for low temperature based solution processing. Nevertheless, still there are remaining challenges of better charge carrier dynamics and defects minimization in HPSCs. Overcoming these challenges are very important to achieve high power conversion efficiency (PCE) practically. Aiming this, a novel organic halide salt called Phenylhydrazinium Iodide (PHAI) is introduced as additive with CH3NH …
Vicarious Methodologies To Assess And Improve The Quality Of The Optical Remote Sensing Images: A Critical Review, Sakib Kabir
Vicarious Methodologies To Assess And Improve The Quality Of The Optical Remote Sensing Images: A Critical Review, Sakib Kabir
Electronic Theses and Dissertations
Over the past decade, number of optical Earth observing satellites performing remote sensing has increased substantially, dramatically increasing the capability to monitor the Earth. The quantity of remote sensing satellite increase is primarily driven by improved technology, miniaturization of components, reduced manufacturing, and launch cost. These satellites often lack on-board calibrators that a large satellite utilizes to ensure high quality (e.g., radiometric, geometric, spatial quality, etc.) scientific measurement. To address this issue, this work presents “best” vicarious image quality assessment and improvement techniques for those kinds of optical satellites which lacks on-board calibration system. In this article, image quality categories …
Surface And Structure Engineering For Next Generation Lithium Metal Batteries, Ke Chen
Surface And Structure Engineering For Next Generation Lithium Metal Batteries, Ke Chen
Electronic Theses and Dissertations
Lithium (Li) metal has been considered as one of the most promising anode materials to replace conventional graphite for Li-ion battery due to its high theoretical capacity (3860 mAh g-1) and low electrochemical potential (-3.04 V vs standard hydrogen electrode). However, it still faces some problems such as unstable solid electrolyte interphase (SEI), uncontrolled Li dendrites growth, and infinite volume change during battery charging/discharging. To develop a stable and low-cost Li metal anode for next-generation Li metal battery, in this dissertation, we have made efforts to understand and solve these problems in two aspects, by introducing an artificial SEI and …
Energy Efficiency And Alternative Energy Generation In Existing Buildings On A College Campus, Sameer Keshavan
Energy Efficiency And Alternative Energy Generation In Existing Buildings On A College Campus, Sameer Keshavan
Electronic Theses and Dissertations
Universities and colleges are looking for ways to be sustainable and save energy costs on their campuses. In terms of raw dollars, “America’s colleges and universities spend almost $7 billion a year in energy and utilities”. Campus buildings consume more than 80% of the energy utilized by the universities and it is often used in the form of electricity for lighting, ventilation, air-conditioning, and natural gas for heating. By reducing energy use, it will translate into cost savings that could be utilized towards student programs, facility improvements, and other university initiatives. This can be achieved by energy conservation efforts and …
Particle Injection Simulation On Cold Spray Technology, Christopher A. Santini
Particle Injection Simulation On Cold Spray Technology, Christopher A. Santini
Electronic Theses and Dissertations
Computational Fluid Dynamics (CFD) is a useful tool when it comes to research in the fields of Aerodynamics, Turbomachinery, and is used in several other fields of research. CFD is an important tool in the engineering industry because it allows for understanding and evaluation of a new design, this can lead to advancements in developing a more efficient and effective design. This tool helps in the understanding of the flow phenomena and how it can interact with its surroundings. One of the main reasons in the use of CFD is to reduce the cost of testing a design by running …
Lithium-Ion Battery And Beyond: Oxygen Vacancy Creation In Tungsten Trioxide And Surface Modification Of Lithium Metal, Rajesh Pathak
Lithium-Ion Battery And Beyond: Oxygen Vacancy Creation In Tungsten Trioxide And Surface Modification Of Lithium Metal, Rajesh Pathak
Electronic Theses and Dissertations
The graphite-based anode material has a low theoretical specific capacity of 371 mAh g-1. The transitional metal oxides (TMOs) are considered a better choice owing to their relatively higher specific capacity. Among TMOs, tungsten trioxide (WO3) is considered promising due to a higher specific capacity of 693 mAh g-1, low cost, mechanically stable, and eco-friendly. It has been a challenge to utilize the TMOs as anode materials as they suffer from poor electronic conductivity and large electrode volume expansion during discharge/charge cycles. In our first project, we demonstrate a unique self-recovery of capacity in reduced WO3 by the incorporation of …
Optimization-Based Fast-Frequency Support In Low Inertia Power Systems, Ujjwol Tamrakar
Optimization-Based Fast-Frequency Support In Low Inertia Power Systems, Ujjwol Tamrakar
Electronic Theses and Dissertations
The future electrical energy demand will largely be met by non-synchronous renewable energy sources (RESs) in the form of photovoltaics and wind energy. The lack of inertial response from these non-synchronous, inverter-based generation in microgrids makes the system vulnerable to large rate-of-change-of-frequency (ROCOF) and frequency excursions. This can trigger under frequency load shedding and cause cascaded outages which may ultimately lead to total blackouts. To limit the ROCOF and the frequency excursions, fast-frequency support can be provided through appropriate control of energy storage systems (ESSs). For proper deployment of such fast-frequency control strategies, accurate information regarding the inertial response of …
Monitoring Methods And Performance Assessment For Implementation Of Low Impact Development Practices In South Dakota, Farhana Akhter
Monitoring Methods And Performance Assessment For Implementation Of Low Impact Development Practices In South Dakota, Farhana Akhter
Electronic Theses and Dissertations
Urbanization increases impervious surface area, which changes the hydrology of a watershed. Impervious surfaces prevent the infiltration of stormwater water into the ground surface which results in a higher volume of stormwater runoff and higher peak flow. Low Impact Development (LID) practices help to restore pre-development hydrology by increasing infiltration, evaporation, and transpiration. While the hydrologic performance of LID practices has been studied extensively and continues to be an area of active research, there has been very limited monitoring or demonstration of LID practices in South Dakota. In addition, municipalities in South Dakota and elsewhere that implement new LID practices …
Internet Of Things Based Wireless Sensor Network And Advanced Machine Learning Models For Precision Agriculture, Ahmed El-Magrous
Internet Of Things Based Wireless Sensor Network And Advanced Machine Learning Models For Precision Agriculture, Ahmed El-Magrous
Electronic Theses and Dissertations
Recent studies assumed that the world population would reach 10.3 billion by 2070. This will require more land for housing; simultaneously resulting in a loss of land for agricultural purposes. However, the new generations also need food, and the lack of new agrarian land is a critical reason that leads researchers and producers to improve daily agriculture practices by using precision agriculture concepts and technologies to increase yield and crop quality. This work represents the design, development, and testing of a customizable and cost-effective Weather-Soil Sensor Station (W-SSS) for use in Precision Agriculture based on high accuracy sensors, wireless communication, …
A Series-Elastic Robot For Back-Pain Rehabilitation, Elhussein Shata
A Series-Elastic Robot For Back-Pain Rehabilitation, Elhussein Shata
Electronic Theses and Dissertations
Robotics research has been broadly expanding into various fields during the past decades. It is widely spread and best known for solving many technical necessities in different fields. With the rise of the industrial revolution, it upgraded many factories to use industrial robots to prevent the human operator from dangerous and hazardous tasks. The rapid development of application fields and their complexity have inspired researchers in the robotics community to find innovative solutions to meet the new desired requirements of the field. Currently, the creation of new needs outside the traditional industrial robots are demanding robots to attend to the …
Data Driven Synthetic Load Modeling For Smart City Energy Management Studies, Fernando Bereta Dos Reis
Data Driven Synthetic Load Modeling For Smart City Energy Management Studies, Fernando Bereta Dos Reis
Electronic Theses and Dissertations
The primary aim of this dissertation is to provide synthetic residential load models with granular level information on the customers having information about the appliances that constitute each individual residential customer through time. The synthetic load model is capable of being widely utilized by the power system research community since only publicly available data is utilized for its generation. This gives researcher’s access to how the synthetic load was made and also how accurate the model is in representing real power system regions. As the title of the dissertation suggests, the synthetic residential load models are intended for smart city …
Investigating The Feasibility, Procedures, Benefits, And Challenges Of Using Mixed Reality In Construction Projects: An Electrical Layout Case Study, Samjhana Shakya
Investigating The Feasibility, Procedures, Benefits, And Challenges Of Using Mixed Reality In Construction Projects: An Electrical Layout Case Study, Samjhana Shakya
Electronic Theses and Dissertations
Innovation and development of technology in the field of visualization have affected the way of designing, presenting, and interacting of the design in the architecture and construction industry. With the invention of the Mixed Reality (MR) device, HoloLens, 2016, many researchers and construction practitioners have marched their research of implementing the MR application in the construction industry. However, there is scarce research on the MR application review, the workflow of using BIM models with the HoloLens, its implementation in the site, and the benefits and challenges of the MR application in the construction site. The gap in the research needs …
Characterization Of Directed Energy Deposition Additively Manufactured Grcop-42 Alloy, Scott Landes
Characterization Of Directed Energy Deposition Additively Manufactured Grcop-42 Alloy, Scott Landes
Electronic Theses and Dissertations
GRCop is an alloy family constructed of copper, chromium, and niobium and was developed by NASA for high heat flux applications. The first of its kind, GRCop-84, was specifically designed for the environments seen by channel cooled main combustion chamber liners. To further increase thermal conductivity while maintaining material strength characteristics, the percentage of alloying elements were cut in half and GRCop- 42 was developed. In recent years, NASA has successfully additively manufactured GRCop with comparable material characteristics to wrought GRCop using a Laser Powder Bed Fusion (L-PBF) process. Benefits of this process include fabrication of intricate cooling channels as …
Pig Pose Estimation Based On Extracted Data Of Mask R-Cnn With Vgg Neural Network For Classifications, Sang Kwan Lee
Pig Pose Estimation Based On Extracted Data Of Mask R-Cnn With Vgg Neural Network For Classifications, Sang Kwan Lee
Electronic Theses and Dissertations
This paper proposes a pig pose estimation operating with Region Proposal Network (RPN) of Mask Region based Convolutional Neural Network (Mask R-CNN) and Visual Geometry Group (VGG) Neural Network (NN). Object pose estimations generates from the associations of different key points. Key points could be explained as specific location of an object such as different joints of a human body or joints of different object. Hourglass network is one of a NN delivering key points of an object. Associating the different key points with the hourglass network results could be represented as instance-level detection [3]. However, the instance-level detection shows …
Kernel-Controlled Dqn Based Cnn Pruning For Model Compression And Acceleration, Romancha Khatri
Kernel-Controlled Dqn Based Cnn Pruning For Model Compression And Acceleration, Romancha Khatri
Electronic Theses and Dissertations
Apart from the accuracy, the size of convolutional neural networks (CNN) models is another principal factor for facilitating the deployment of models on memory, power and budget constrained devices. However, conventional model compression techniques require human experts to setup parameters to explore the design space which is suboptimal and time consuming. Various pruning techniques are implemented to gain compression, trading off speed and accuracy. Given a CNN model [11], we propose an automated deep reinforcement learning [9] based model compression technique that can effectively turned off kernels on each layer by observing its significance on decision making. By observing accuracy, …