Digital Commons Network^™

Histogram Matching For Camera Pose Neighbor Selection, Parris K. Egbert, Bryan S. Morse, Kevin L. Steele

Faculty Publications

A prerequisite to calibrated camera pose estimation is the construction of a camera neighborhood adjacency graph, a connected graph defining the pose neighbors of the camera set. Pose neighbors to a camera C are images containing sufficient overlap in image content with the image from C that they can be used to correctly estimate the pose of C using structure-from-motion techniques. In a video stream, the camera neighborhood adjacency graph is often a simple connected path; frame poses are only estimated relative to their immediate neighbors. We propose a novel method to build more complex camera adjacency graphs that are …

Multilearner Based Recursive Supervised Training, Kiruthika Ramanathan, Sheng Uei Guan, Laxmi R. Iyer

Research Collection School Of Computing and Information Systems

In supervised learning, most single solution neural networks such as constructive backpropagation give good results when used with some datasets but not with others. Others such as probabilistic neural networks (PNN) fit a curve to perfection but need to be manually tuned in the case of noisy data. Recursive percentage based hybrid pattern training (RPHP) overcomes this problem by recursively training subsets of the data, thereby using several neural networks. MultiLearner based recursive training (MLRT) is an extension of this approach, where a combination of existing and new learners are used and subsets are trained using the weak learner which …

Exploiting Domain Structure For Named Entity Recognition, Jing Jiang, Chengxiang Zhai

Research Collection School Of Computing and Information Systems

Named Entity Recognition (NER) is a fundamental task in text mining and natural language understanding. Current approaches to NER (mostly based on supervised learning) perform well on domains similar to the training domain, but they tend to adapt poorly to slightly different domains. We present several strategies for exploiting the domain structure in the training data to learn a more robust named entity recognizer that can perform well on a new domain. First, we propose a simple yet effective way to automatically rank features based on their generalizabilities across domains. We then train a classifier with strong emphasis on the …

Batch Mode Active Learning And Its Applications To Medical Image Classification, Steven C. H. Hoi, Rong Jin, Jianke Zhu, Michael R. Lyu

Research Collection School Of Computing and Information Systems

The goal of active learning is to select the most informative examples for manual labeling. Most of the previous studies in active learning have focused on selecting a single unlabeled example in each iteration. This could be inefficient since the classification model has to be retrained for every labeled example. In this paper, we present a framework for "batch mode active learning" that applies the Fisher information matrix to select a number of informative examples simultaneously. The key computational challenge is how to efficiently identify the subset of unlabeled examples that can result in the largest reduction in the Fisher …

Learning Distance Metrics With Contextual Constraints For Image Retrieval, Steven C. H. Hoi, Wei Liu, Michael R. Lyu, Wei-Ying Ma

Research Collection School Of Computing and Information Systems

Relevant Component Analysis (RCA) has been proposed for learning distance metrics with contextual constraints for image retrieval. However, RCA has two important disadvantages. One is the lack of exploiting negative constraints which can also be informative, and the other is its incapability of capturing complex nonlinear relationships between data instances with the contextual information. In this paper, we propose two algorithms to overcome these two disadvantages, i.e., Discriminative Component Analysis (DCA) and Kernel DCA. Compared with other complicated methods for distance metric learning, our algorithms are rather simple to understand and very easy to solve. We evaluate the performance of …

Nanomanipulation Using Atomic Force Microscope With Drift Compensation, Qinmin Yang, Jagannathan Sarangapani

Electrical and Computer Engineering Faculty Research & Creative Works

This paper proposes an atomic force microscope (AFM) based force controller to push nanoparticles on the substrates since it is tedious for human. A block phase correlation-based algorithm is embedded into the controller for compensating the thermal drift during nanomanipulation. Further, a neural network (NN) is employed to approximate the unknown nanoparticle and substrate contact dynamics including the roughness effects. Using the NN-based adaptive force controller the task of pushing nanoparticles is demonstrated. Finally, using the Lyapunov-based stability analysis, the uniform ultimately boundedness (UUB) of the closed-loop signals is demonstrated

Adaptive Neural Network Control And Wireless Sensor Network Based Localization For Uav Formation, H. Wu, Jagannathan Sarangapani

Electrical and Computer Engineering Faculty Research & Creative Works

We consider a team of unmanned aerial vehicles (UAV's) equipped with sensors and motes for wireless communication for the task of navigating to a desired location in a formation. First a neural network (NN)-based control scheme is presented that allows the UAVs to track a desired position and orientation with reference to the neighboring UAVs or obstacles in the environment. Second, we discuss a graph theory-based scheme for discovery, localization and cooperative control. The purpose of the NN cooperative controller is to achieve and maintain the desired formation shape in the presence of unmodeled dynamics and bounded unknown disturbances. Numerical …

Structural Auditing Methodologies For Controlled Terminologies, Hua Min

Dissertations

Several auditing methodologies for large controlled terminologies are developed. These are applied to the Unified Medical Language System XXXX and the National Cancer Institute Thesaurus (NCIT). Structural auditing methodologies are based on the structural aspects such as IS-A hierarchy relationships groups of concepts assigned to semantic types and groups of relationships defined for concepts. Structurally uniform groups of concepts tend to be semantically uniform. Structural auditing methodologies focus on concepts with unlikely or rare configuration. These concepts have a high likelihood for errors.

One of the methodologies is based on comparing hierarchical relationships between the META and SN, two major …

Image Vectorization, Brian L. Price

Theses and Dissertations

We present a new technique for creating an editable vector graphic from an object in a raster image. Object selection is performed interactively in subsecond time by calling graph cut with each mouse movement. A renderable mesh is then computed automatically for the selected object and each of its (sub)objects by (1) generating a coarse object mesh; (2) performing recursive graph cut segmentation and hierarchical ordering of subobjects; (3) applying error-driven mesh refinement to each (sub)object. The result is a fully layered object hierarchy that facilitates object-level editing without leaving holes. Object-based vectorization compares favorably with current approaches in the …

Temporal Data Mining In A Dynamic Feature Space, Brent K. Wenerstrom

Theses and Dissertations

Many interesting real-world applications for temporal data mining are hindered by concept drift. One particular form of concept drift is characterized by changes to the underlying feature space. Seemingly little has been done to address this issue. This thesis presents FAE, an incremental ensemble approach to mining data subject to concept drift. FAE achieves better accuracies over four large datasets when compared with a similar incremental learning algorithm.

Trust Negotiation For Open Database Access Control, Paul A. Porter

Theses and Dissertations

Hippocratic databases are designed to protect the privacy of the individuals whose personal information they contain. This thesis presents a model for providing and enforcing access control in an open Hippocratic database system. Previously unknown individuals can gain access to information in the database by authenticating to roles through trust negotiation. Allowing qualified strangers to access the database increases the usefulness of the system without compromising privacy. This thesis presents the design and implementation of two methods for filtering information from database queries. First, we extend a query modification method for use in an open database system. Second, we introduce …

Relative Clock Drift Rate Based Secure Time Synchronization For Wireless Sensor Networks, Jae Sung Choi

Computer Science and Engineering Theses - Archive

Time synchronization is critical issue to many wireless sensor network applications such that target tracking, TDMA radio scheduling, and secure localization. However, the most of existing time synchronization algorithms in wireless sensor networks did not consider malicious attacks in hostile environments. In this thesis, we propose a Relative Clock Drift Rate Based Secure Time Synchronization (RSTS) schemes to address security problems. RSTS alleviates delay attacks and incorrect time stamp transmissions caused by external or internal malicious attackers. We discuss a simple estimation technique to calculate a relative clock drift rate between sender-receiver clocks, because the estimated relative clock drift rate …

Towards Optimum Play-Out Buffering Delay In Voice Over Ip, Ruchir Pramod Shende

Computer Science and Engineering Theses - Archive

Voice over Internet Protocol (VoIP) or transmission of real-time voice packets over the Internet is slowly emerging as a cost-effective alternative to the traditional Public Switched Telephone Network (PSTN). However, varying end-to-end delay and packet loss, which are inherent in a packet-switched network like the Internet, lead to relatively lower quality of VoIP calls. The call quality can be improved by adaptively adjusting the play-out buffer at the receiver to reduce the impact of the delay and jitter. A standard play-out strategy uses a weighted moving average of the mean and variance of network delay to adaptively set the play-out …

Reducing Price Fluctuation In Continuous Double Auctions Through Pricing Policy And Shout Improvement, Jinzhong Niu, Kai Cai, Simon Parsons, Elizabeth Sklar

Publications and Research

Auction mechanism design is a subfield of game theory dedicated to manipulating the rules of an auction so as to achieve specific goals. The complexity of the dynamics of auctions, especially continuous double auctions, makes it difficult to apply the usual analytic game-theoretic methods. This paper takes an experimental approach, searching a parameterized space of possible auction types, and presents a new pricing policy for continuous double auctions. The paper further demonstrates how this policy, together with a shout improvement rule, helps to reduce the fluctuation of transaction prices in auctions involving agents with minimum intelligence while keeping the overall …

Concept-Based Search Using Parallel Query Expansion, Rahul Rajiv Joshi

Computer Science and Engineering Theses - Archive

We address the problem of irrelevant results for short queries on Web search engines. Short queries fail to provide sufficient context to disambiguate possible meanings associated with the search terms resulting in a set of irrelevant pages that the user has to filter through navigation and sometimes examination. First, we predict the potential concept topics, which are the domains for the search terms. This prediction is based on word occurrences and relationships observed in the various domains (categories) of a corpus. Next, we expand the search terms in each of the predicted domains in parallel. We then submit separate queries, …

Ubca: A Utility Based Clustering Architecture For Peer-To-Peer Networks, Brent Lagesse

Computer Science and Engineering Theses - Archive

Use of the Peer-to-Peer (P2P) architecture has recently spread in popularity. File sharing and ad hoc networks have contributed to the architecture's usage. P2P generates new challenges in scalability, fairness, and quality of service. Current solutions tend to fall into two main areas: incentives and system design. Incentive-based approaches appeal to the self-interested nature of peers by requiring service to the system in order to access resources. System design includes distributed hash tables and graph-theoretical based designs which have seen some success, but also result in new problems. We introduce a Utility-Based Clustering Architecture, UBCA, designed to address scalability, fairness, …

Alternative Architectures For Improving Document Readability, Ankur Bora

Computer Science and Engineering Theses - Archive

A number of tools have been developed to improve the readability of documents. These tools assist users to make changes which result in an easy to read document. Most of these tools were developed for users who work independently with little interaction with other users. However, in recent years, because of the proliferation of Internet, there has been increased collaboration between users. A user may need to share his document repositories with others so that both can benefit. Another set of users may work with information from a specific domain. These domains may be located in different geographical areas. Some …

Anytime Top-K Queries On Exact And Fuzzy Data, Bhushan P. Chaudhari

Computer Science and Engineering Theses - Archive

Top-k queries on large multi-attribute data sets are fundamental operations in information retrieval and ranking applications. In this thesis, we initiate research on the anytime behavior of top-k algorithms on exact and fuzzy data. In particular given specific topk algorithms we are interested in studying their progress towards identification of the correct result at any point of the algorithms' execution. We adopt a probabilistic approach where we seek to report at any point the scores of the top-k results the algorithm has identified, as well as associate a confidence with this prediction. Such functionality can be a valuable asset when …

Towards A Stronger Peer-To-Peer Anonymous System, Arjun R. Nambiar

Computer Science and Engineering Theses - Archive

Anonymous communications systems on the Internet provides protection against eavesdroppers and others that seek to link users with their communications. These systems have many important applications in areas such as law enforcement, intelligence gathering, business privacy, anonymous publishing, and personal privacy. Currently deployed systems rely on a relatively small set of advertised servers to forward messages for the user. These systems can suffer from scalability problems, with potentially large bandwidth and system overhead costs, and the servers themselves can be targets of direct attacks. Peer-to-peer anonymous communications systems, such as Tarzan[1] and MorphMix [2], have been proposed as a way …

Investigation Of Techniques To Increase The Scalability Of Graph-Based Data Mining Algorithms., Srilatha Inavolu

Computer Science and Engineering Theses - Archive

Frequent subgraph pattern recognition and graph-based relational learning have been an emerging area of data mining research with scientific and commercial applications. At the kernel of these algorithms are the computationally-expensive graph and subgraph isomorphism tests. The graph isomorphism problem consists in deciding whether two graphs are isomorphic i.e., whether there is a one-one mapping between the vertices of the two graphs that respects the edge connections. Many graphs will be depicted quite differently but in actuality have the same inherent structure. This leads to the isomorphism problem. The graph isomorphism problem belongs to the class of NP problems and …

Multiple Object Tracking Using Particle Filters, Hwangryol Ryu

Computer Science and Engineering Theses - Archive

We describe a novel extension to the Particle filter algorithm for tracking multiple objects. The recently proposed algorithms and the variants for multiple object tacking algorithms estimate multi-modal posterior distributions that potentially represent the multiple peaks (i.e., multiple tracked objects). However, the speci¯c state representation does not demonstrate birth, death and more importantly partial/complete occlusion of the objects. Furthermore, the weakness of the Particle Filter such that the representation may increasingly bias the posterior density estimates toward objects with dominant likelihood makes the multiple object tracking algorithms more di±cult. To circumvent a sample depletion problem and maintain the computational complexity …

Supervised Learning From Embedded Subgraphs, Joseph T. Potts

Computer Science and Engineering Dissertations - Archive

We develop a machine learning algorithm which learns rules for classification from training examples in a graph representation. However, unlike most other such algorithms which use one graph for each example, ours allows all of the training examples to be in a single, connected graph. We employ the Minimum Description Length principle to produce a novel performance metric for judging the value of a learned classification. We implement the algorithm by extending the Subdue graph-based learning system. Finally, we demonstrate the use of the new system in two different domains, earth science and homeland security.

Semantic Analytics Visualization, Leonidas Deligiannidis, Amit P. Sheth, Boanerges Aleman-Meza

Kno.e.sis Publications

In this paper we present a new tool for semantic analytics through 3D visualization called “Semantic Analytics Visualization” (SAV). It has the capability for visualizing ontologies and meta-data including annotated web-documents, images, and digital media such as audio and video clips in a synthetic three-dimensional semi-immersive environment. More importantly, SAV supports visual semantic analytics, whereby an analyst can interactively investigate complex relationships between heterogeneous information. The tool is built using Virtual Reality technology which makes SAV a highly interactive system. The backend of SAV consists of a Semantic Analytics system that supports query processing and semantic association discovery. Using a …

On In-Network Synopsis Join Processing For Sensor Networks, Hai Yu, Ee Peng Lim, Jun Zhang

Research Collection School Of Computing and Information Systems

The emergence of sensor networks enables applications that deploy sensors to collaboratively monitor environment and process data collected. In some scenarios, we are interested in using join queries to correlate data stored in different regions of a sensor network, where the data volume is large, making it prohibitive to transmit all data to a central server for joining. In this paper, we present an in-network synopsis join strategy for evaluating join queries in sensor networks with communication efficiency. In this strategy, we prune data that do not contribute to the join results in the early stage of the join processing, …

Large-Scale Text Categorization By Batch Mode Active Learning, Steven C. H. Hoi, Rong Jin, Michael R. Lyu

Research Collection School Of Computing and Information Systems

Large-scale text categorization is an important research topic for Web data mining. One of the challenges in large-scale text categorization is how to reduce the human efforts in labeling text documents for building reliable classification models. In the past, there have been many studies on applying active learning methods to automatic text categorization, which try to select the most informative documents for labeling manually. Most of these studies focused on selecting a single unlabeled document in each iteration. As a result, the text categorization model has to be retrained after each labeled document is solicited. In this paper, we present …

Real-Time Non-Rigid Shape Recovery Via Active Appearance Models For Augmented Reality, Jianke Zhu, Steven C. H. Hoi, Michael R. Lyu

Research Collection School Of Computing and Information Systems

One main challenge in Augmented Reality (AR) applications is to keep track of video objects with their movement, orientation, size, and position accurately. This poses a challenging task to recover nonrigid shape and global pose in real-time AR applications. This paper proposes a novel two-stage scheme for online non-rigid shape recovery toward AR applications using Active Appearance Models (AAMs). First, we construct 3D shape models from AAMs offline, which do not involve processing of the 3D scan data. Based on the computed 3D shape models, we propose an efficient online algorithm to estimate both 3D pose and non-rigid shape parameters …

Mining Rdf Metadata For Generalized Association Rules: Knowledge Discovery In The Semantic Web Era, Tao Jiang, Ah-Hwee Tan

Research Collection School Of Computing and Information Systems

In this paper, we present a novel frequent generalized pattern mining algorithm, called GP-Close, for mining generalized associations from RDF metadata. To solve the over-generalization problem encountered by existing methods, GP-Close employs the notion of emphgeneralization closure for systematic over-generalization reduction.

Gestalt-Based Feature Similarity Measure In Trademark Database, Hui Jiang, Chong-Wah Ngo, Hung-Khoon Tan

Research Collection School Of Computing and Information Systems

Motivated by the studies in Gestalt principle, this paper describes a novel approach on the adaptive selection of visual features for trademark retrieval. We consider five kinds of visual saliencies: symmetry, continuity, proximity, parallelism and closure property. The first saliency is based on Zernike moments, while the others are modeled by geometric elements extracted illusively as a whole from a trademark. Given a query trademark, we adaptively determine the features appropriate for retrieval by investigating its visual saliencies. We show that in most cases, either geometric or symmetric features can give us good enough accuracy. To measure the similarity of …

Clip-Based Similarity Measure For Query-Dependent Clip Retrieval And Video Summarization, Yuxin Peng, Chong-Wah Ngo

Research Collection School Of Computing and Information Systems

This paper proposes a new approach and algorithm for the similarity measure of video clips. The similarity is mainly based on two bipartite graph matching algorithms: maximum matching (MM) and optimal matching (OM). MM is able to rapidly filter irrelevant video clips, while OM is capable of ranking the similarity of clips according to visual and granularity factors. We apply the similarity measure for two tasks: retrieval and summarization. In video retrieval, a hierarchical retrieval framework is constructed based on MM and OM. The validity of the framework is theoretically proved and empirically verified on a video database of 21 …

Efficient Querying And Resource Management Using Distributed Presence Information In Converged Networks, Dipanjan Chakraborty, Koustuv Dasgupta, Archan Misra

Research Collection School Of Computing and Information Systems

Next-generation converged networks shall deliver many innovative services over the standardized SIPbased IMS signaling infrastructure. Several such services exploit the joint presence information of a consumer, i.e. SIP entity requesting a service, and a vendor, i.e. SIP resource providing a service. Presence information is a collection of contextual attributes (e.g. location, availability, reputation), some of which change dynamically. Moreover, this collective presence information is distributed across multiple presence servers. While performing query matching based on joint presence information, a server usually routes each query to a locally available resource. However, skews in the spatio-temporal distribution of queries and resources may …