Search
Now showing items 1-20 of 124
Integrate template matching and statistical modeling for continuous speech recognition
(University of Missouri--Columbia, 2011)
In this dissertation, a novel approach of integrating template matching with statistical modeling is proposed to improve continuous speech recognition. Commonly used Hidden Markov Models (HMMs) are ineffective in modeling details of speech temporal...
Statistical model-based methods for observation selection in wireless sensor networks and for feature selection in classification
(University of Missouri--Columbia, 2012)
We apply statistical model-based approaches to address temporal and spatial observation selection challenges in wireless sensor networks. For temporal observation selection, we present an improved version of VoIDP algorithm that is the first optimal...
Analysis of host-pathogen interactions via clustering, statistical analysis, and data visualization
(University of Missouri--Columbia, 2015)
Infectious diseases are caused by a variety of agents: viruses, bacteria, parasites, or even proteins. Using existing state-of-the-art methods and tools I developed myself, I studied aspects of infectious agents. To find ...
Shape reconstruction from volumetric images
(University of Missouri--Columbia, 2008)
initialization in a 2-D grayscale image, a PDE is proposed to evolve an explicit model based on the interior statistics of the previously identified region. The algorithm is adapted to volumetric diffusion tensor MRI to extract the thalamus and thalamus nuclei...
Statistical optimization of acoustic models for large vocabulary speech recognition
(University of Missouri--Columbia, 2006)
This dissertation investigates optimization of acoustic models in speech recognition. Two new optimization methods are proposed for phonetic decision tree (PDT) search and Hidden Markov modeling (HMM)-- the knowledge-based ...
A protocol for simultaneous real time playback and full quality storage of streaming media
(University of Missouri--Columbia, 2010)
This research introduces the new problem of simultaneous streaming of a single media bitstream to multiple devices with different Quality of Service (QoS) requirements. In particular, the research addresses simultaneous ...
Bioinformatics methods for protein identification using peptide mass fingerprinting data
(University of Missouri--Columbia, 2009)
) are not normalized. Therefore, the ranking based on raw scores may be biased. To address the above issue, we have developed a statistical model to evaluate the confidence of the raw score and to improve the ranking of proteins for identification. (3) Software...
Genome scale meta analysis of microarrays for biological inferences
(University of Missouri--Columbia, 2009)
In this present era of high-throughput technologies, meta-analysis is being widely used to integrate multiple similar high throughput studies. Here we propose a novel framework for applying meta-analysis techniques on ...
Deepnet : an extensible data acquisition and curation framework supporting computer vision deep learning research
(University of Missouri--Columbia, 2018)
public data sets, the framework includes open source research data points which are found during the course of curation and quality assurance of data while utilizing the framework. These features are of particular value since there is a real possibility...
A new adaptive framework for collaborative filtering prediction
(University of Missouri--Columbia, 2008)
-based collaborative filtering (CF). Existing CF techniques work well on dense data but poorly on sparse data. To address this weakness, we propose to use z-scores instead of explicit ratings and introduce a mechanism that adaptively combines global statistics...
Statistical inference in wireless sensor and mobile networks
(University of Missouri--Columbia, 2010)
) prediction. We formulate them as spatial/temporal statistical inference problems and develop efficient centralized and decentralized solution methods. In the problem of data authentication, we aim at providing an energy efficient means for data authentication...
Language modeling for automatic speech recognition in telehealth
(University of Missouri--Columbia, 2005)
Standard statistic n-gram language models play a critical and indispensable role in automatic speech recognition (ASR) applications. Though helpful to ASR, it suffers from a practical problem when lacking sufficient in-domain training data that come...
Sylview: a visualization system for forest management
(University of Missouri--Columbia, 2006)
In the field of forestry, the output of forest growth models provide a wealth of detailed information that can often be difficult to analyze and perceive due to presentation either as plain text summary tables or static ...
Decoder-learning based distributed source coding for high-efficiency, low-cost and secure multimedia communications
(University of Missouri--Columbia, 2008)
Conventional multimedia compression leverages the source statistics at the encoder side. This is not suitable for some emerging applications such as wireless sensor networks, where the encoders usually have limited functionalities and power supplies...
Development of web services in academic environment
(University of Missouri--Columbia, 2019)
) Grant Dashboard and COE Request System as examples. The COE Grant Dashboard is an ASP.NET website used to manage and statistically analyze records of research awards, expenditures, and proposals in the College of Engineering. The COE Request System is a...
Detecting targeted data poisoning attacks on deep neural networks
(University of Missouri--Columbia, 2022)
Deep neural networks (DNNs) are widely used for various facial image-recognition purposes, including facial recognition and subsequent authentication, and the detection of altered facial images. Unfortunately, due to their ...
Knowledge-driven multidimensional indexing for content-based information retrieval
(University of Missouri--Columbia, 2008)
Statistical (EBS) k-d tree and the Entropy Balanced Bitmap (EBB) tree for knowledge-driven indexing of media. These indexing structures exploit domain knowledge present in a media database through analysis of the statistical entropy of the database...
Estimation and tracking of elder activity levels for health event prediction
(University of Missouri--Columbia, 2009)
often be predicted, and thus, be avoided or reduced in severity. Statistics on activities of daily living and activity level over an extended period of time provide important data for functional assessment and health prediction. However, persistent...
"Bring-your-own" plug-in management for next-generation science gateway applications
(University of Missouri--Columbia, 2020)
applications. The OnTimeRecommend Adviser features a variety of recommender modules to help novice/expert users with knowledge discovery through data sources such as e.g., publications, funding records, cloud templates and Jupyter notebooks. Based...
Domain-concept mining : an efficient on-demand data mining approach
(University of Missouri--Columbia, 2008)
by pairing each attribute with its values. Criteria may include under-represented groups as well as spatial, temporal, and incremental dimensionalities. Then, a statistical power analysis is utilized to determine if multiple criteria of the same attribute...