ASU Electronic Theses and Dissertations
This collection includes most of the ASU Theses and Dissertations from 2011 to present. ASU Theses and Dissertations are available in downloadable PDF format; however, a small percentage of items are under embargo. Information about the dissertations/theses includes degree information, committee members, an abstract, supporting data or media.
In addition to the electronic theses found in the ASU Digital Repository, ASU Theses and Dissertations can be found in the ASU Library Catalog.
Dissertations and Theses granted by Arizona State University are archived and made available through a joint effort of the ASU Graduate College and the ASU Libraries. For more information or questions about this collection contact or visit the Digital Repository ETD Library Guide or contact the ASU Graduate College at gradformat@asu.edu.
Displaying 1 - 10 of 137
Description
Production from a high pressure gas well at a high production-rate encounters the risk of operating near the choking condition for a compressible flow in porous media. The unbounded gas pressure gradient near the point of choking, which is located near the wellbore, generates an effective tensile stress on the porous rock frame. This tensile stress almost always exceeds the tensile strength of the rock and it causes a tensile failure of the rock, leading to wellbore instability. In a porous rock, not all pores are choked at the same flow rate, and when just one pore is choked, the flow through the entire porous medium should be considered choked as the gas pressure gradient at the point of choking becomes singular. This thesis investigates the choking condition for compressible gas flow in a single microscopic pore. Quasi-one-dimensional analysis and axisymmetric numerical simulations of compressible gas flow in a pore scale varicose tube with a number of bumps are carried out, and the local Mach number and pressure along the tube are computed for the flow near choking condition. The effects of tube length, inlet-to-outlet pressure ratio, the number of bumps and the amplitude of the bumps on the choking condition are obtained. These critical values provide guidance for avoiding the choking condition in practice.
ContributorsYuan, Jing (Author) / Chen, Kangping (Thesis advisor) / Wang, Liping (Committee member) / Huang, Huei-Ping (Committee member) / Arizona State University (Publisher)
Created2013
Description
Tesla turbo-machinery offers a robust, easily manufactured, extremely versatile prime mover with inherent capabilities making it perhaps the best, if not the only, solution for certain niche applications. The goal of this thesis is not to optimize the performance of the Tesla turbine, but to compare its performance with various working fluids. Theoretical and experimental analyses of a turbine-generator assembly utilizing compressed air, saturated steam and water as the working fluids were performed and are presented in this work. A brief background and explanation of the technology is provided along with potential applications. A theoretical thermodynamic analysis is outlined, resulting in turbine and rotor efficiencies, power outputs and Reynolds numbers calculated for the turbine for various combinations of working fluids and inlet nozzles. The results indicate the turbine is capable of achieving a turbine efficiency of 31.17 ± 3.61% and an estimated rotor efficiency 95 ± 9.32%. These efficiencies are promising considering the numerous losses still present in the current design. Calculation of the Reynolds number provided some capability to determine the flow behavior and how that behavior impacts the performance and efficiency of the Tesla turbine. It was determined that turbulence in the flow is essential to achieving high power outputs and high efficiency. Although the efficiency, after peaking, begins to slightly taper off as the flow becomes increasingly turbulent, the power output maintains a steady linear increase.
ContributorsPeshlakai, Aaron (Author) / Phelan, Patrick (Thesis advisor) / Trimble, Steve (Committee member) / Wang, Liping (Committee member) / Arizona State University (Publisher)
Created2012
Description
Nanostructured materials show signicant enhancement in the thermoelectric g-
ure of merit (zT) due to quantum connement eects. Improving the eciency of
thermoelectric devices allows for the development of better, more economical waste
heat recovery systems. Such systems may be used as bottoming or co-generation
cycles in conjunction with conventional power cycles to recover some of the wasted
heat. Thermal conductivity measurement systems are an important part of the char-
acterization processes of thermoelectric materials. These systems must possess the
capability of accurately measuring the thermal conductivity of both bulk and thin-lm
samples at dierent ambient temperatures.
This paper discusses the construction, validation, and improvement of a thermal
conductivity measurement platform based on the 3-Omega technique. Room temperature
measurements of thermal conductivity done on control samples with known properties
such as undoped bulk silicon (Si), bulk gallium arsenide (GaAs), and silicon dioxide
(SiO2) thin lms yielded 150 W=m􀀀K, 50 W=m􀀀K, and 1:46 W=m􀀀K respectively.
These quantities were all within 8% of literature values. In addition, the thermal
conductivity of bulk SiO2 was measured as a function of temperature in a Helium-
4 cryostat from 75K to 250K. The results showed good agreement with literature
values that all fell within the error range of each measurement. The uncertainty in
the measurements ranged from 19% at 75K to 30% at 250K. Finally, the system
was used to measure the room temperature thermal conductivity of a nanocomposite
composed of cadmium selenide, CdSe, nanocrystals in an indium selenide, In2Se3,
matrix as a function of the concentration of In2Se3. The observed trend was in
qualitative agreement with the expected behavior.
i
ure of merit (zT) due to quantum connement eects. Improving the eciency of
thermoelectric devices allows for the development of better, more economical waste
heat recovery systems. Such systems may be used as bottoming or co-generation
cycles in conjunction with conventional power cycles to recover some of the wasted
heat. Thermal conductivity measurement systems are an important part of the char-
acterization processes of thermoelectric materials. These systems must possess the
capability of accurately measuring the thermal conductivity of both bulk and thin-lm
samples at dierent ambient temperatures.
This paper discusses the construction, validation, and improvement of a thermal
conductivity measurement platform based on the 3-Omega technique. Room temperature
measurements of thermal conductivity done on control samples with known properties
such as undoped bulk silicon (Si), bulk gallium arsenide (GaAs), and silicon dioxide
(SiO2) thin lms yielded 150 W=m􀀀K, 50 W=m􀀀K, and 1:46 W=m􀀀K respectively.
These quantities were all within 8% of literature values. In addition, the thermal
conductivity of bulk SiO2 was measured as a function of temperature in a Helium-
4 cryostat from 75K to 250K. The results showed good agreement with literature
values that all fell within the error range of each measurement. The uncertainty in
the measurements ranged from 19% at 75K to 30% at 250K. Finally, the system
was used to measure the room temperature thermal conductivity of a nanocomposite
composed of cadmium selenide, CdSe, nanocrystals in an indium selenide, In2Se3,
matrix as a function of the concentration of In2Se3. The observed trend was in
qualitative agreement with the expected behavior.
i
ContributorsJaber, Abbas (Author) / Wang, Robert (Thesis advisor) / Wang, Liping (Committee member) / Rykaczewski, Konrad (Committee member) / Arizona State University (Publisher)
Created2014
Description
The rapid progress of solution-phase synthesis has led colloidal nanocrystals one of the most versatile nanoscale materials, provided opportunities to tailor material's properties, and boosted related technological innovations. Colloidal nanocrystal-based materials have been demonstrated success in a variety of applications, such as LEDs, electronics, solar cells and thermoelectrics. In each of these applications, the thermal transport property plays a big role. An undesirable temperature rise due to inefficient heat dissipation could lead to deleterious effects on devices' performance and lifetime. Hence, the first project is focused on investigating the thermal transport in colloidal nanocrystal solids. This study answers the question that how the molecular structure of nanocrystals affect the thermal transport, and provides insights for future device designs. In particular, PbS nanocrystals is used as a monitoring system, and the core diameter, ligand length and ligand binding group are systematically varied to study the corresponding effect on thermal transport.
Next, a fundamental study is presented on the phase stability and solid-liquid transformation of metallic (In, Sn and Bi) colloidal nanocrystals. Although the phase change of nanoparticles has been a long-standing research topic, the melting behavior of colloidal nanocrytstals is largely unexplored. In addition, this study is of practical importance to nanocrystal-based applications that operate at elevated temperatures. Embedding colloidal nanocrystals into thermally-stable polymer matrices allows preserving nanocrystal size throughout melt-freeze cycles, and therefore enabling observation of stable melting features. Size-dependent melting temperature, melting enthalpy and melting entropy have all been measured and discussed.
In the next two chapters, focus has been switched to developing colloidal nanocrystal-based phase change composites for thermal energy storage applications. In Chapter 4, a polymer matrix phase change nanocomposite has been created. In this composite, the melting temperature and energy density could be independently controlled by tuning nanocrystal diameter and volume fractions. In Chapter 5, a solution-phase synthesis on metal matrix-metal nanocrytal composite is presented. This approach enables excellent morphological control over nanocrystals and demonstrated a phase change composite with a thermal conductivity 2 - 3 orders of magnitude greater than typical phase change materials, such as organics and molten salts.
Next, a fundamental study is presented on the phase stability and solid-liquid transformation of metallic (In, Sn and Bi) colloidal nanocrystals. Although the phase change of nanoparticles has been a long-standing research topic, the melting behavior of colloidal nanocrytstals is largely unexplored. In addition, this study is of practical importance to nanocrystal-based applications that operate at elevated temperatures. Embedding colloidal nanocrystals into thermally-stable polymer matrices allows preserving nanocrystal size throughout melt-freeze cycles, and therefore enabling observation of stable melting features. Size-dependent melting temperature, melting enthalpy and melting entropy have all been measured and discussed.
In the next two chapters, focus has been switched to developing colloidal nanocrystal-based phase change composites for thermal energy storage applications. In Chapter 4, a polymer matrix phase change nanocomposite has been created. In this composite, the melting temperature and energy density could be independently controlled by tuning nanocrystal diameter and volume fractions. In Chapter 5, a solution-phase synthesis on metal matrix-metal nanocrytal composite is presented. This approach enables excellent morphological control over nanocrystals and demonstrated a phase change composite with a thermal conductivity 2 - 3 orders of magnitude greater than typical phase change materials, such as organics and molten salts.
ContributorsLiu, Minglu (Author) / Wang, Robert Y (Thesis advisor) / Wang, Liping (Committee member) / Rykaczewski, Konrad (Committee member) / Phelan, Patrick (Committee member) / Dai, Lenore (Committee member) / Arizona State University (Publisher)
Created2015
Description
Nanoparticle suspensions, popularly termed “nanofluids,” have been extensively investigated for their thermal and radiative properties. Such work has generated great controversy, although it is arguably accepted today that the presence of nanoparticles rarely leads to useful enhancements in either thermal conductivity or convective heat transfer. On the other hand, there are still examples of unanticipated enhancements to some properties, such as the reported specific heat of molten salt-based nanofluids and the critical heat flux. Another largely overlooked example is the apparent effect of nanoparticles on the effective latent heat of vaporization (hfg) of aqueous nanofluids. A previous study focused on molecular dynamics (MD) modeling supplemented with limited experimental data to suggest that hfg increases with increasing nanoparticle concentration.
Here, this research extends that exploratory work in an effort to determine if hfg of aqueous nanofluids can be manipulated, i.e., increased or decreased, by the addition of graphite or silver nanoparticles. Our results to date indicate that hfg can be substantially impacted, by up to ± 30% depending on the type of nanoparticle. Moreover, this dissertation reports further experiments with changing surface area based on volume fraction (0.005% to 2%) and various nanoparticle sizes to investigate the mechanisms for hfg modification in aqueous graphite and silver nanofluids. This research also investigates thermophysical properties, i.e., density and surface tension in aqueous nanofluids to support the experimental results of hfg based on the Clausius - Clapeyron equation. This theoretical investigation agrees well with the experimental results. Furthermore, this research investigates the hfg change of aqueous nanofluids with nanoscale studies in terms of melting of silver nanoparticles and hydrophobic interactions of graphite nanofluid. As a result, the entropy change due to those mechanisms could be a main cause of the changes of hfg in silver and graphite nanofluids.
Finally, applying the latent heat results of graphite and silver nanofluids to an actual solar thermal system to identify enhanced performance with a Rankine cycle is suggested to show that the tunable latent heat of vaporization in nanofluilds could be beneficial for real-world solar thermal applications with improved efficiency.
Here, this research extends that exploratory work in an effort to determine if hfg of aqueous nanofluids can be manipulated, i.e., increased or decreased, by the addition of graphite or silver nanoparticles. Our results to date indicate that hfg can be substantially impacted, by up to ± 30% depending on the type of nanoparticle. Moreover, this dissertation reports further experiments with changing surface area based on volume fraction (0.005% to 2%) and various nanoparticle sizes to investigate the mechanisms for hfg modification in aqueous graphite and silver nanofluids. This research also investigates thermophysical properties, i.e., density and surface tension in aqueous nanofluids to support the experimental results of hfg based on the Clausius - Clapeyron equation. This theoretical investigation agrees well with the experimental results. Furthermore, this research investigates the hfg change of aqueous nanofluids with nanoscale studies in terms of melting of silver nanoparticles and hydrophobic interactions of graphite nanofluid. As a result, the entropy change due to those mechanisms could be a main cause of the changes of hfg in silver and graphite nanofluids.
Finally, applying the latent heat results of graphite and silver nanofluids to an actual solar thermal system to identify enhanced performance with a Rankine cycle is suggested to show that the tunable latent heat of vaporization in nanofluilds could be beneficial for real-world solar thermal applications with improved efficiency.
ContributorsLee, Soochan (Author) / Phelan, Patrick E (Thesis advisor) / Wu, Carole-Jean (Thesis advisor) / Wang, Robert (Committee member) / Wang, Liping (Committee member) / Taylor, Robert A. (Committee member) / Prasher, Ravi (Committee member) / Arizona State University (Publisher)
Created2015
Description
In a collaborative environment where multiple robots and human beings are expected
to collaborate to perform a task, it becomes essential for a robot to be aware of multiple
agents working in its work environment. A robot must also learn to adapt to
different agents in the workspace and conduct its interaction based on the presence
of these agents. A theoretical framework was introduced which performs interaction
learning from demonstrations in a two-agent work environment, and it is called
Interaction Primitives.
This document is an in-depth description of the new state of the art Python
Framework for Interaction Primitives between two agents in a single as well as multiple
task work environment and extension of the original framework in a work environment
with multiple agents doing a single task. The original theory of Interaction
Primitives has been extended to create a framework which will capture correlation
between more than two agents while performing a single task. The new state of the
art Python framework is an intuitive, generic, easy to install and easy to use python
library which can be applied to use the Interaction Primitives framework in a work
environment. This library was tested in simulated environments and controlled laboratory
environment. The results and benchmarks of this library are available in the
related sections of this document.
to collaborate to perform a task, it becomes essential for a robot to be aware of multiple
agents working in its work environment. A robot must also learn to adapt to
different agents in the workspace and conduct its interaction based on the presence
of these agents. A theoretical framework was introduced which performs interaction
learning from demonstrations in a two-agent work environment, and it is called
Interaction Primitives.
This document is an in-depth description of the new state of the art Python
Framework for Interaction Primitives between two agents in a single as well as multiple
task work environment and extension of the original framework in a work environment
with multiple agents doing a single task. The original theory of Interaction
Primitives has been extended to create a framework which will capture correlation
between more than two agents while performing a single task. The new state of the
art Python framework is an intuitive, generic, easy to install and easy to use python
library which can be applied to use the Interaction Primitives framework in a work
environment. This library was tested in simulated environments and controlled laboratory
environment. The results and benchmarks of this library are available in the
related sections of this document.
ContributorsKumar, Ashish, M.S (Author) / Amor, Hani Ben (Thesis advisor) / Zhang, Yu (Committee member) / Yang, Yezhou (Committee member) / Arizona State University (Publisher)
Created2017
Description
Computer Vision as a eld has gone through signicant changes in the last decade.
The eld has seen tremendous success in designing learning systems with hand-crafted
features and in using representation learning to extract better features. In this dissertation
some novel approaches to representation learning and task learning are studied.
Multiple-instance learning which is generalization of supervised learning, is one
example of task learning that is discussed. In particular, a novel non-parametric k-
NN-based multiple-instance learning is proposed, which is shown to outperform other
existing approaches. This solution is applied to a diabetic retinopathy pathology
detection problem eectively.
In cases of representation learning, generality of neural features are investigated
rst. This investigation leads to some critical understanding and results in feature
generality among datasets. The possibility of learning from a mentor network instead
of from labels is then investigated. Distillation of dark knowledge is used to eciently
mentor a small network from a pre-trained large mentor network. These studies help
in understanding representation learning with smaller and compressed networks.
The eld has seen tremendous success in designing learning systems with hand-crafted
features and in using representation learning to extract better features. In this dissertation
some novel approaches to representation learning and task learning are studied.
Multiple-instance learning which is generalization of supervised learning, is one
example of task learning that is discussed. In particular, a novel non-parametric k-
NN-based multiple-instance learning is proposed, which is shown to outperform other
existing approaches. This solution is applied to a diabetic retinopathy pathology
detection problem eectively.
In cases of representation learning, generality of neural features are investigated
rst. This investigation leads to some critical understanding and results in feature
generality among datasets. The possibility of learning from a mentor network instead
of from labels is then investigated. Distillation of dark knowledge is used to eciently
mentor a small network from a pre-trained large mentor network. These studies help
in understanding representation learning with smaller and compressed networks.
ContributorsVenkatesan, Ragav (Author) / Li, Baoxin (Thesis advisor) / Turaga, Pavan (Committee member) / Yang, Yezhou (Committee member) / Davulcu, Hasan (Committee member) / Arizona State University (Publisher)
Created2017
Description
With the rise of the Big Data Era, an exponential amount of network data is being generated at an unprecedented rate across a wide-range of high impact micro and macro areas of research---from protein interaction to social networks. The critical challenge is translating this large scale network data into actionable information.
A key task in the data translation is the analysis of network connectivity via marked nodes---the primary focus of our research. We have developed a framework for analyzing network connectivity via marked nodes in large scale graphs, utilizing novel algorithms in three interrelated areas: (1) analysis of a single seed node via it’s ego-centric network (AttriPart algorithm); (2) pathway identification between two seed nodes (K-Simple Shortest Paths Multithreaded and Search Reduced (KSSPR) algorithm); and (3) tree detection, defining the interaction between three or more seed nodes (Shortest Path MST algorithm).
In an effort to address both fundamental and applied research issues, we have developed the LocalForcasting algorithm to explore how network connectivity analysis can be applied to local community evolution and recommender systems. The goal is to apply the LocalForecasting algorithm to various domains---e.g., friend suggestions in social networks or future collaboration in co-authorship networks. This algorithm utilizes link prediction in combination with the AttriPart algorithm to predict future connections in local graph partitions.
Results show that our proposed AttriPart algorithm finds up to 1.6x denser local partitions, while running approximately 43x faster than traditional local partitioning techniques (PageRank-Nibble). In addition, our LocalForecasting algorithm demonstrates a significant improvement in the number of nodes and edges correctly predicted over baseline methods. Furthermore, results for the KSSPR algorithm demonstrate a speed-up of up to 2.5x the standard k-simple shortest paths algorithm.
A key task in the data translation is the analysis of network connectivity via marked nodes---the primary focus of our research. We have developed a framework for analyzing network connectivity via marked nodes in large scale graphs, utilizing novel algorithms in three interrelated areas: (1) analysis of a single seed node via it’s ego-centric network (AttriPart algorithm); (2) pathway identification between two seed nodes (K-Simple Shortest Paths Multithreaded and Search Reduced (KSSPR) algorithm); and (3) tree detection, defining the interaction between three or more seed nodes (Shortest Path MST algorithm).
In an effort to address both fundamental and applied research issues, we have developed the LocalForcasting algorithm to explore how network connectivity analysis can be applied to local community evolution and recommender systems. The goal is to apply the LocalForecasting algorithm to various domains---e.g., friend suggestions in social networks or future collaboration in co-authorship networks. This algorithm utilizes link prediction in combination with the AttriPart algorithm to predict future connections in local graph partitions.
Results show that our proposed AttriPart algorithm finds up to 1.6x denser local partitions, while running approximately 43x faster than traditional local partitioning techniques (PageRank-Nibble). In addition, our LocalForecasting algorithm demonstrates a significant improvement in the number of nodes and edges correctly predicted over baseline methods. Furthermore, results for the KSSPR algorithm demonstrate a speed-up of up to 2.5x the standard k-simple shortest paths algorithm.
ContributorsFreitas, Scott (Author) / Tong, Hanghang (Thesis advisor) / Maciejewski, Ross (Committee member) / Yang, Yezhou (Committee member) / Arizona State University (Publisher)
Created2018
Description
This dissertation will investigate two of the most promising high-capacity anode
materials for lithium-based batteries: silicon (Si) and metal lithium (Li). It will focus on
studying the mechanical behaviors of the two materials during charge and discharge and
understanding how these mechanical behaviors may affect their electrochemical
performance.
In the first part, amorphous Si anode will be studied. Despite many existing studies
on silicon (Si) anodes for lithium ion batteries (LIBs), many essential questions still exist
on compound formation, composition, and properties. Here it is shown that some
previously accepted findings do not truthfully reflect the actual lithiation mechanisms in
realistic battery configurations. Furthermore the correlation between structure and
mechanical properties in these materials has not been properly established. Here, a rigorous
and thorough study is performed to comprehensively understand the electrochemical
reaction mechanisms of amorphous-Si (a-Si) in a realistic LIB configuration. In-depth
microstructural characterization was performed and correlations were established between
Li-Si composition, volumetric expansion, and modulus/hardness. It is found that the
lithiation process of a-Si in a real battery setup is a single-phase reaction rather than the
accepted two-phase reaction obtained from in-situ TEM experiments. The findings in this
dissertation establish a reference to quantitatively explain many key metrics for lithiated a
Si as anodes in real LIBs, and can be used to rationally design a-Si based high-performance
LIBs guided by high-fidelity modeling and simulations.
In the second part, Li metal anode will be investigated. Problems related to dendrite
growth on lithium metal anodes such as capacity loss and short circuit present major
barriers to the next-generation high-energy-density batteries. The development of
successful mitigation strategies is impeded by the incomplete understanding of the Li
dendrite growth mechanisms. Here the enabling role of plating residual stress in dendrite
initiation through novel experiments of Li electrodeposition on soft substrates is confirmed,
and the observations is explained with a stress-driven dendrite growth model. Dendrite
growth is mitigated on such soft substrates through surface-wrinkling-induced stress
relaxation in deposited Li film. It is demonstrated that this new dendrite mitigation
mechanism can be utilized synergistically with other existing approaches in the form of
three-dimensional (3D) soft scaffolds for Li plating, which achieves superior coulombic
efficiency over conventional hard copper current collectors under large current density.
materials for lithium-based batteries: silicon (Si) and metal lithium (Li). It will focus on
studying the mechanical behaviors of the two materials during charge and discharge and
understanding how these mechanical behaviors may affect their electrochemical
performance.
In the first part, amorphous Si anode will be studied. Despite many existing studies
on silicon (Si) anodes for lithium ion batteries (LIBs), many essential questions still exist
on compound formation, composition, and properties. Here it is shown that some
previously accepted findings do not truthfully reflect the actual lithiation mechanisms in
realistic battery configurations. Furthermore the correlation between structure and
mechanical properties in these materials has not been properly established. Here, a rigorous
and thorough study is performed to comprehensively understand the electrochemical
reaction mechanisms of amorphous-Si (a-Si) in a realistic LIB configuration. In-depth
microstructural characterization was performed and correlations were established between
Li-Si composition, volumetric expansion, and modulus/hardness. It is found that the
lithiation process of a-Si in a real battery setup is a single-phase reaction rather than the
accepted two-phase reaction obtained from in-situ TEM experiments. The findings in this
dissertation establish a reference to quantitatively explain many key metrics for lithiated a
Si as anodes in real LIBs, and can be used to rationally design a-Si based high-performance
LIBs guided by high-fidelity modeling and simulations.
In the second part, Li metal anode will be investigated. Problems related to dendrite
growth on lithium metal anodes such as capacity loss and short circuit present major
barriers to the next-generation high-energy-density batteries. The development of
successful mitigation strategies is impeded by the incomplete understanding of the Li
dendrite growth mechanisms. Here the enabling role of plating residual stress in dendrite
initiation through novel experiments of Li electrodeposition on soft substrates is confirmed,
and the observations is explained with a stress-driven dendrite growth model. Dendrite
growth is mitigated on such soft substrates through surface-wrinkling-induced stress
relaxation in deposited Li film. It is demonstrated that this new dendrite mitigation
mechanism can be utilized synergistically with other existing approaches in the form of
three-dimensional (3D) soft scaffolds for Li plating, which achieves superior coulombic
efficiency over conventional hard copper current collectors under large current density.
ContributorsWang, Xu (Author) / Jiang, Hanqing (Thesis advisor) / Yu, Hongbin (Thesis advisor) / Chan, Candace (Committee member) / Wang, Liping (Committee member) / Qiong, Nian (Committee member) / Arizona State University (Publisher)
Created2018
Description
The performance of most of the visual computing tasks depends on the quality of the features extracted from the raw data. Insightful feature representation increases the performance of many learning algorithms by exposing the underlying explanatory factors of the output for the unobserved input. A good representation should also handle anomalies in the data such as missing samples and noisy input caused by the undesired, external factors of variation. It should also reduce the data redundancy. Over the years, many feature extraction processes have been invented to produce good representations of raw images and videos.
The feature extraction processes can be categorized into three groups. The first group contains processes that are hand-crafted for a specific task. Hand-engineering features requires the knowledge of domain experts and manual labor. However, the feature extraction process is interpretable and explainable. Next group contains the latent-feature extraction processes. While the original feature lies in a high-dimensional space, the relevant factors for a task often lie on a lower dimensional manifold. The latent-feature extraction employs hidden variables to expose the underlying data properties that cannot be directly measured from the input. Latent features seek a specific structure such as sparsity or low-rank into the derived representation through sophisticated optimization techniques. The last category is that of deep features. These are obtained by passing raw input data with minimal pre-processing through a deep network. Its parameters are computed by iteratively minimizing a task-based loss.
In this dissertation, I present four pieces of work where I create and learn suitable data representations. The first task employs hand-crafted features to perform clinically-relevant retrieval of diabetic retinopathy images. The second task uses latent features to perform content-adaptive image enhancement. The third task ranks a pair of images based on their aestheticism. The goal of the last task is to capture localized image artifacts in small datasets with patch-level labels. For both these tasks, I propose novel deep architectures and show significant improvement over the previous state-of-art approaches. A suitable combination of feature representations augmented with an appropriate learning approach can increase performance for most visual computing tasks.
The feature extraction processes can be categorized into three groups. The first group contains processes that are hand-crafted for a specific task. Hand-engineering features requires the knowledge of domain experts and manual labor. However, the feature extraction process is interpretable and explainable. Next group contains the latent-feature extraction processes. While the original feature lies in a high-dimensional space, the relevant factors for a task often lie on a lower dimensional manifold. The latent-feature extraction employs hidden variables to expose the underlying data properties that cannot be directly measured from the input. Latent features seek a specific structure such as sparsity or low-rank into the derived representation through sophisticated optimization techniques. The last category is that of deep features. These are obtained by passing raw input data with minimal pre-processing through a deep network. Its parameters are computed by iteratively minimizing a task-based loss.
In this dissertation, I present four pieces of work where I create and learn suitable data representations. The first task employs hand-crafted features to perform clinically-relevant retrieval of diabetic retinopathy images. The second task uses latent features to perform content-adaptive image enhancement. The third task ranks a pair of images based on their aestheticism. The goal of the last task is to capture localized image artifacts in small datasets with patch-level labels. For both these tasks, I propose novel deep architectures and show significant improvement over the previous state-of-art approaches. A suitable combination of feature representations augmented with an appropriate learning approach can increase performance for most visual computing tasks.
ContributorsChandakkar, Parag Shridhar (Author) / Li, Baoxin (Thesis advisor) / Yang, Yezhou (Committee member) / Turaga, Pavan (Committee member) / Davulcu, Hasan (Committee member) / Arizona State University (Publisher)
Created2017