Theses and Dissertations
This collection includes both ASU Theses and Dissertations, submitted by graduate students, and the Barrett, Honors College theses submitted by undergraduate students.
Displaying 1 - 10 of 15
Filtering by
- Creators: Li, Baoxin
Description
We solve the problem of activity verification in the context of sustainability. Activity verification is the process of proving the user assertions pertaining to a certain activity performed by the user. Our motivation lies in incentivizing the user for engaging in sustainable activities like taking public transport or recycling. Such incentivization schemes require the system to verify the claim made by the user. The system verifies these claims by analyzing the supporting evidence captured by the user while performing the activity. The proliferation of portable smart-phones in the past few years has provided us with a ubiquitous and relatively cheap platform, having multiple sensors like accelerometer, gyroscope, microphone etc. to capture this evidence data in-situ. In this research, we investigate the supervised and semi-supervised learning techniques for activity verification. Both these techniques make use the data set constructed using the evidence submitted by the user. Supervised learning makes use of annotated evidence data to build a function to predict the class labels of the unlabeled data points. The evidence data captured can be either unimodal or multimodal in nature. We use the accelerometer data as evidence for transportation mode verification and image data as evidence for recycling verification. After training the system, we achieve maximum accuracy of 94% when classifying the transport mode and 81% when detecting recycle activity. In the case of recycle verification, we could improve the classification accuracy by asking the user for more evidence. We present some techniques to ask the user for the next best piece of evidence that maximizes the probability of classification. Using these techniques for detecting recycle activity, the accuracy increases to 93%. The major disadvantage of using supervised models is that it requires extensive annotated training data, which expensive to collect. Due to the limited training data, we look at the graph based inductive semi-supervised learning methods to propagate the labels among the unlabeled samples. In the semi-supervised approach, we represent each instance in the data set as a node in the graph. Since it is a complete graph, edges interconnect these nodes, with each edge having some weight representing the similarity between the points. We propagate the labels in this graph, based on the proximity of the data points to the labeled nodes. We estimate the performance of these algorithms by measuring how close the probability distribution of the data after label propagation is to the probability distribution of the ground truth data. Since labeling has a cost associated with it, in this thesis we propose two algorithms that help us in selecting minimum number of labeled points to propagate the labels accurately. Our proposed algorithm achieves a maximum of 73% increase in performance when compared to the baseline algorithm.
ContributorsDesai, Vaishnav (Author) / Sundaram, Hari (Thesis advisor) / Li, Baoxin (Thesis advisor) / Turaga, Pavan (Committee member) / Arizona State University (Publisher)
Created2013
Description
Software has a great impact on the energy efficiency of any computing system--it can manage the components of a system efficiently or inefficiently. The impact of software is amplified in the context of a wearable computing system used for activity recognition. The design space this platform opens up is immense and encompasses sensors, feature calculations, activity classification algorithms, sleep schedules, and transmission protocols. Design choices in each of these areas impact energy use, overall accuracy, and usefulness of the system. This thesis explores methods software can influence the trade-off between energy consumption and system accuracy. In general the more energy a system consumes the more accurate will be. We explore how finding the transitions between human activities is able to reduce the energy consumption of such systems without reducing much accuracy. We introduce the Log-likelihood Ratio Test as a method to detect transitions, and explore how choices of sensor, feature calculations, and parameters concerning time segmentation affect the accuracy of this method. We discovered an approximate 5X increase in energy efficiency could be achieved with only a 5% decrease in accuracy. We also address how a system's sleep mode, in which the processor enters a low-power state and sensors are turned off, affects a wearable computing platform that does activity recognition. We discuss the energy trade-offs in each stage of the activity recognition process. We find that careful analysis of these parameters can result in great increases in energy efficiency if small compromises in overall accuracy can be tolerated. We call this the ``Great Compromise.'' We found a 6X increase in efficiency with a 7% decrease in accuracy. We then consider how wireless transmission of data affects the overall energy efficiency of a wearable computing platform. We find that design decisions such as feature calculations and grouping size have a great impact on the energy consumption of the system because of the amount of data that is stored and transmitted. For example, storing and transmitting vector-based features such as FFT or DCT do not compress the signal and would use more energy than storing and transmitting the raw signal. The effect of grouping size on energy consumption depends on the feature. For scalar features energy consumption is proportional in the inverse of grouping size, so it's reduced as grouping size goes up. For features that depend on the grouping size, such as FFT, energy increases with the logarithm of grouping size, so energy consumption increases slowly as grouping size increases. We find that compressing data through activity classification and transition detection significantly reduces energy consumption and that the energy consumed for the classification overhead is negligible compared to the energy savings from data compression. We provide mathematical models of energy usage and data generation, and test our ideas using a mobile computing platform, the Texas Instruments Chronos watch.
ContributorsBoyd, Jeffrey Michael (Author) / Sundaram, Hari (Thesis advisor) / Li, Baoxin (Thesis advisor) / Shrivastava, Aviral (Committee member) / Turaga, Pavan (Committee member) / Arizona State University (Publisher)
Created2014
Description
Many learning models have been proposed for various tasks in visual computing. Popular examples include hidden Markov models and support vector machines. Recently, sparse-representation-based learning methods have attracted a lot of attention in the computer vision field, largely because of their impressive performance in many applications. In the literature, many of such sparse learning methods focus on designing or application of some learning techniques for certain feature space without much explicit consideration on possible interaction between the underlying semantics of the visual data and the employed learning technique. Rich semantic information in most visual data, if properly incorporated into algorithm design, should help achieving improved performance while delivering intuitive interpretation of the algorithmic outcomes. My study addresses the problem of how to explicitly consider the semantic information of the visual data in the sparse learning algorithms. In this work, we identify four problems which are of great importance and broad interest to the community. Specifically, a novel approach is proposed to incorporate label information to learn a dictionary which is not only reconstructive but also discriminative; considering the formation process of face images, a novel image decomposition approach for an ensemble of correlated images is proposed, where a subspace is built from the decomposition and applied to face recognition; based on the observation that, the foreground (or salient) objects are sparse in input domain and the background is sparse in frequency domain, a novel and efficient spatio-temporal saliency detection algorithm is proposed to identify the salient regions in video; and a novel hidden Markov model learning approach is proposed by utilizing a sparse set of pairwise comparisons among the data, which is easier to obtain and more meaningful, consistent than tradition labels, in many scenarios, e.g., evaluating motion skills in surgical simulations. In those four problems, different types of semantic information are modeled and incorporated in designing sparse learning algorithms for the corresponding visual computing tasks. Several real world applications are selected to demonstrate the effectiveness of the proposed methods, including, face recognition, spatio-temporal saliency detection, abnormality detection, spatio-temporal interest point detection, motion analysis and emotion recognition. In those applications, data of different modalities are involved, ranging from audio signal, image to video. Experiments on large scale real world data with comparisons to state-of-art methods confirm the proposed approaches deliver salient advantages, showing adding those semantic information dramatically improve the performances of the general sparse learning methods.
ContributorsZhang, Qiang (Author) / Li, Baoxin (Thesis advisor) / Turaga, Pavan (Committee member) / Wang, Yalin (Committee member) / Ye, Jieping (Committee member) / Arizona State University (Publisher)
Created2014
Description
In the sport of competitive water skiing, the skill of a human boat driver can affect athletic performance. Driver influence is not necessarily inhibitive to skiers, however, it reduces the fairness and credibility of the sport overall. In response to the stated problem, this thesis proposes a vision-based real-time control system designed specifically for tournament waterski boats. The challenges addressed in this thesis include: one, the segmentation of floating objects in frame sequences captured by a moving camera, two, the identification of segmented objects which fit a predefined model, and three, the accurate and fast estimation of camera position and orientation from coplanar point correspondences. This thesis discusses current ideas and proposes new methods for the three challenges mentioned. In the end, a working prototype is produced.
ContributorsWalker, Collin (Author) / Li, Baoxin (Thesis advisor) / Turaga, Pavan (Committee member) / Claveau, David (Committee member) / Arizona State University (Publisher)
Created2014
Description
Diabetic retinopathy (DR) is a common cause of blindness occurring due to prolonged presence of diabetes. The risk of developing DR or having the disease progress is increasing over time. Despite advances in diabetes care over the years, DR remains a vision-threatening complication and one of the leading causes of blindness among American adults. Recent studies have shown that diagnosis based on digital retinal imaging has potential benefits over traditional face-to-face evaluation. Yet there is a dearth of computer-based systems that can match the level of performance achieved by ophthalmologists. This thesis takes a fresh perspective in developing a computer-based system aimed at improving diagnosis of DR images. These images are categorized into three classes according to their severity level. The proposed approach explores effective methods to classify new images and retrieve clinically-relevant images from a database with prior diagnosis information associated with them. Retrieval provides a novel way to utilize the vast knowledge in the archives of previously-diagnosed DR images and thereby improve a clinician's performance while classification can safely reduce the burden on DR screening programs and possibly achieve higher detection accuracy than human experts. To solve the three-class retrieval and classification problem, the approach uses a multi-class multiple-instance medical image retrieval framework that makes use of spectrally tuned color correlogram and steerable Gaussian filter response features. The results show better retrieval and classification performances than prior-art methods and are also observed to be of clinical and visual relevance.
ContributorsChandakkar, Parag Shridhar (Author) / Li, Baoxin (Thesis advisor) / Turaga, Pavan (Committee member) / Frakes, David (Committee member) / Arizona State University (Publisher)
Created2012
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
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
Description
Mixture of experts is a machine learning ensemble approach that consists of individual models that are trained to be ``experts'' on subsets of the data, and a gating network that provides weights to output a combination of the expert predictions. Mixture of experts models do not currently see wide use due to difficulty in training diverse experts and high computational requirements. This work presents modifications of the mixture of experts formulation that use domain knowledge to improve training, and incorporate parameter sharing among experts to reduce computational requirements.
First, this work presents an application of mixture of experts models for quality robust visual recognition. First it is shown that human subjects outperform deep neural networks on classification of distorted images, and then propose a model, MixQualNet, that is more robust to distortions. The proposed model consists of ``experts'' that are trained on a particular type of image distortion. The final output of the model is a weighted sum of the expert models, where the weights are determined by a separate gating network. The proposed model also incorporates weight sharing to reduce the number of parameters, as well as increase performance.
Second, an application of mixture of experts to predict visual saliency is presented. A computational saliency model attempts to predict where humans will look in an image. In the proposed model, each expert network is trained to predict saliency for a set of closely related images. The final saliency map is computed as a weighted mixture of the expert networks' outputs, with weights determined by a separate gating network. The proposed model achieves better performance than several other visual saliency models and a baseline non-mixture model.
Finally, this work introduces a saliency model that is a weighted mixture of models trained for different levels of saliency. Levels of saliency include high saliency, which corresponds to regions where almost all subjects look, and low saliency, which corresponds to regions where some, but not all subjects look. The weighted mixture shows improved performance compared with baseline models because of the diversity of the individual model predictions.
First, this work presents an application of mixture of experts models for quality robust visual recognition. First it is shown that human subjects outperform deep neural networks on classification of distorted images, and then propose a model, MixQualNet, that is more robust to distortions. The proposed model consists of ``experts'' that are trained on a particular type of image distortion. The final output of the model is a weighted sum of the expert models, where the weights are determined by a separate gating network. The proposed model also incorporates weight sharing to reduce the number of parameters, as well as increase performance.
Second, an application of mixture of experts to predict visual saliency is presented. A computational saliency model attempts to predict where humans will look in an image. In the proposed model, each expert network is trained to predict saliency for a set of closely related images. The final saliency map is computed as a weighted mixture of the expert networks' outputs, with weights determined by a separate gating network. The proposed model achieves better performance than several other visual saliency models and a baseline non-mixture model.
Finally, this work introduces a saliency model that is a weighted mixture of models trained for different levels of saliency. Levels of saliency include high saliency, which corresponds to regions where almost all subjects look, and low saliency, which corresponds to regions where some, but not all subjects look. The weighted mixture shows improved performance compared with baseline models because of the diversity of the individual model predictions.
ContributorsDodge, Samuel Fuller (Author) / Karam, Lina (Thesis advisor) / Jayasuriya, Suren (Committee member) / Li, Baoxin (Committee member) / Turaga, Pavan (Committee member) / Arizona State University (Publisher)
Created2018
Description
The rapid growth of social media in recent years provides a large amount of user-generated visual objects, e.g., images and videos. Advanced semantic understanding approaches on such visual objects are desired to better serve applications such as human-machine interaction, image retrieval, etc. Semantic visual attributes have been proposed and utilized in multiple visual computing tasks to bridge the so-called "semantic gap" between extractable low-level feature representations and high-level semantic understanding of the visual objects.
Despite years of research, there are still some unsolved problems on semantic attribute learning. First, real-world applications usually involve hundreds of attributes which requires great effort to acquire sufficient amount of labeled data for model learning. Second, existing attribute learning work for visual objects focuses primarily on images, with semantic analysis on videos left largely unexplored.
In this dissertation I conduct innovative research and propose novel approaches to tackling the aforementioned problems. In particular, I propose robust and accurate learning frameworks on both attribute ranking and prediction by exploring the correlation among multiple attributes and utilizing various types of label information. Furthermore, I propose a video-based skill coaching framework by extending attribute learning to the video domain for robust motion skill analysis. Experiments on various types of applications and datasets and comparisons with multiple state-of-the-art baseline approaches confirm that my proposed approaches can achieve significant performance improvements for the general attribute learning problem.
Despite years of research, there are still some unsolved problems on semantic attribute learning. First, real-world applications usually involve hundreds of attributes which requires great effort to acquire sufficient amount of labeled data for model learning. Second, existing attribute learning work for visual objects focuses primarily on images, with semantic analysis on videos left largely unexplored.
In this dissertation I conduct innovative research and propose novel approaches to tackling the aforementioned problems. In particular, I propose robust and accurate learning frameworks on both attribute ranking and prediction by exploring the correlation among multiple attributes and utilizing various types of label information. Furthermore, I propose a video-based skill coaching framework by extending attribute learning to the video domain for robust motion skill analysis. Experiments on various types of applications and datasets and comparisons with multiple state-of-the-art baseline approaches confirm that my proposed approaches can achieve significant performance improvements for the general attribute learning problem.
ContributorsChen, Lin (Author) / Li, Baoxin (Thesis advisor) / Turaga, Pavan (Committee member) / Wang, Yalin (Committee member) / Liu, Huan (Committee member) / Arizona State University (Publisher)
Created2016
Description
High-level inference tasks in video applications such as recognition, video retrieval, and zero-shot classification have become an active research area in recent years. One fundamental requirement for such applications is to extract high-quality features that maintain high-level information in the videos.
Many video feature extraction algorithms have been purposed, such as STIP, HOG3D, and Dense Trajectories. These algorithms are often referred to as “handcrafted” features as they were deliberately designed based on some reasonable considerations. However, these algorithms may fail when dealing with high-level tasks or complex scene videos. Due to the success of using deep convolution neural networks (CNNs) to extract global representations for static images, researchers have been using similar techniques to tackle video contents. Typical techniques first extract spatial features by processing raw images using deep convolution architectures designed for static image classifications. Then simple average, concatenation or classifier-based fusion/pooling methods are applied to the extracted features. I argue that features extracted in such ways do not acquire enough representative information since videos, unlike images, should be characterized as a temporal sequence of semantically coherent visual contents and thus need to be represented in a manner considering both semantic and spatio-temporal information.
In this thesis, I propose a novel architecture to learn semantic spatio-temporal embedding for videos to support high-level video analysis. The proposed method encodes video spatial and temporal information separately by employing a deep architecture consisting of two channels of convolutional neural networks (capturing appearance and local motion) followed by their corresponding Fully Connected Gated Recurrent Unit (FC-GRU) encoders for capturing longer-term temporal structure of the CNN features. The resultant spatio-temporal representation (a vector) is used to learn a mapping via a Fully Connected Multilayer Perceptron (FC-MLP) to the word2vec semantic embedding space, leading to a semantic interpretation of the video vector that supports high-level analysis. I evaluate the usefulness and effectiveness of this new video representation by conducting experiments on action recognition, zero-shot video classification, and semantic video retrieval (word-to-video) retrieval, using the UCF101 action recognition dataset.
Many video feature extraction algorithms have been purposed, such as STIP, HOG3D, and Dense Trajectories. These algorithms are often referred to as “handcrafted” features as they were deliberately designed based on some reasonable considerations. However, these algorithms may fail when dealing with high-level tasks or complex scene videos. Due to the success of using deep convolution neural networks (CNNs) to extract global representations for static images, researchers have been using similar techniques to tackle video contents. Typical techniques first extract spatial features by processing raw images using deep convolution architectures designed for static image classifications. Then simple average, concatenation or classifier-based fusion/pooling methods are applied to the extracted features. I argue that features extracted in such ways do not acquire enough representative information since videos, unlike images, should be characterized as a temporal sequence of semantically coherent visual contents and thus need to be represented in a manner considering both semantic and spatio-temporal information.
In this thesis, I propose a novel architecture to learn semantic spatio-temporal embedding for videos to support high-level video analysis. The proposed method encodes video spatial and temporal information separately by employing a deep architecture consisting of two channels of convolutional neural networks (capturing appearance and local motion) followed by their corresponding Fully Connected Gated Recurrent Unit (FC-GRU) encoders for capturing longer-term temporal structure of the CNN features. The resultant spatio-temporal representation (a vector) is used to learn a mapping via a Fully Connected Multilayer Perceptron (FC-MLP) to the word2vec semantic embedding space, leading to a semantic interpretation of the video vector that supports high-level analysis. I evaluate the usefulness and effectiveness of this new video representation by conducting experiments on action recognition, zero-shot video classification, and semantic video retrieval (word-to-video) retrieval, using the UCF101 action recognition dataset.
ContributorsHu, Sheng-Hung (Author) / Li, Baoxin (Thesis advisor) / Turaga, Pavan (Committee member) / Liang, Jianming (Committee member) / Tong, Hanghang (Committee member) / Arizona State University (Publisher)
Created2016