Matthew T. Moores

Speechbot: an experimental speech-based search engine for multimedia content on the web

As the Web transforms from a text-only medium into a more multimedia-rich medium, the need arises to perform searches based on the multimedia content. In this paper, we present an audio and video search engine to tackle this problem. The engine uses speech recognition technology to index spoken audio and video files from the World Wide Web (WWW) when no transcriptions are available. If transcriptions (even imperfect ones) are available, we can also take advantage of them to improve the indexing process. Our engine indexes several thousand talk and news radio shows covering a wide range of topics and speaking styles from a selection of public Web sites with multimedia archives. Our Web site is similar in spirit to normal Web search sites; it contains an index, not the actual multimedia content. The audio from these shows suffers in acoustic quality due to bandwidth limitations, coding, compression, and poor acoustic conditions. Our word error rate (WER) results using appropriately trained acoustic models show remarkable resilience to the high compression, although many factors combine to increase the average WERs over standard broadcast news benchmarks. We show that, even if the transcription is inaccurate, we can still achieve good retrieval performance for typical user queries (77.5%).

The recycling endosome protein Rab17 regulates melanocytic filopodia formation and melanosome trafficking

Rab GTPases including Rab27a, Rab38 and Rab32 function in melanosome maturation or trafficking in melanocytes. A screen to identify additional Rabs involved in these processes revealed the localization of GFP‐Rab17 on recycling endosomes (REs) and melanosomes in melanocytic cells. Rab17 mRNA expression is regulated by microphthalmia transcription factor (MITF), a characteristic of known pigmentation genes. Rab17 siRNA knockdown in melanoma cells quantitatively increased melanosome concentration at the cell periphery. Rab17 knockdown did not inhibit melanosome maturation nor movement, but it caused accumulation of melanin inside cells. Double knockdown of Rab17 and Rab27a indicated that Rab17 acts on melanosomes downstream of Rab27a. Filopodia are known to play a role in melanosome transfer, and in Rab17 knockdown cells filopodia formation was inhibited. Furthermore, we show that stimulation of melanoma cells with α‐melanocyte‐stimulating hormone induces filopodia formation, supporting a role for filopodia in melanosome release. Cell stimulation also caused redistribution of REs to the periphery, and knockdown of additional RE‐associated Rabs 11a and 11b produced a similar accumulation of melanosomes and melanin to that seen after loss of Rab17. Our findings reveal new functions for RE and Rab17 in pigmentation through a distal step in the process of melanosome release via filopodia.

Preferential Attachment of Specific Fluorescent Dyes and Dye Labeled DNA Sequences in a Surface Enhanced Raman Scattering Multiplex

A significant advantage of using surface enhanced Raman scattering (SERS) for DNA detection is the capability to detect multiple analytes simultaneously within the one sample. However, as the analytes approach the metallic surface required for SERS, they become more concentrated and previous studies have suggested that different dye labels will have different affinities for the metal surface. Here, the interaction of single stranded DNA labeled with either fluorescein (FAM) or tetramethylrhodamine (TAMRA) with a metal surface, using spermine induced aggregated silver nanoparticles as the SERS substrate, is investigated by analyzing the labels separately and in mixtures. Comparison studies were also undertaken using the dyes in their free isothiocyanate forms, fluorescein isothiocyanate (F-ITC) and tetramethylrhodamine isothiocyanate (TR-ITC). When the two dyes are premixed prior to the addition of nanoparticles, TAMRA exerts a strong masking effect over FAM due to a stronger affinity for the metal surface. When parameters such as order of analyte addition, analysis time, and analyte concentration are investigated, the masking effect of TAMRA is still observed but the extent changes depending on the experimental parameters. By using bootstrap estimation of changes in SERS peak intensity, a greater insight has been achieved into the surface affinity of the two dyes as well as how they interact with each other. It has been shown that the order of addition of the analytes is important and that specific dye related interactions occur, which could greatly affect the observed SERS spectra. SERS has been used successfully for the simultaneous detection of several analytes; however, this work has highlighted the significant factors that must be taken into consideration when planning a multiple analyte assay.

Segmentation of cone-beam CT using a hidden Markov random field with informative priors

Cone-beam computed tomography (CBCT) has enormous potential to improve the accuracy of treatment delivery in image-guided radiotherapy (IGRT). To assist radiotherapists in interpreting these images, we use a Bayesian statistical model to label each voxel according to its tissue type. The rich sources of prior information in IGRT are incorporated into a hidden Markov random field model of the 3D image lattice. Tissue densities in the reference CT scan are estimated using inverse regression and then rescaled to approximate the corresponding CBCT intensity values. The treatment planning contours are combined with published studies of physiological variability to produce a spatial prior distribution for changes in the size, shape and position of the tumour volume and organs at risk. The voxel labels are estimated using iterated conditional modes. The accuracy of the method has been evaluated using 27 CBCT scans of an electron density phantom. The mean voxel-wise misclassification rate was 6.2%, with Dice similarity coefficient of 0.73 for liver, muscle, breast and adipose tissue. By incorporating prior information, we are able to successfully segment CBCT images. This could be a viable approach for automated, online image analysis in radiotherapy.

Accelerating pseudo-marginal MCMC using Gaussian processes

The grouped independence Metropolis–Hastings (GIMH) and Markov chain within Metropolis (MCWM) algorithms are pseudo-marginal methods used to perform Bayesian inference in latent variable models. These methods replace intractable likelihood calculations with unbiased estimates within Markov chain Monte Carlo algorithms. The GIMH method has the posterior of interest as its limiting distribution, but suffers from poor mixing if it is too computationally intensive to obtain high-precision likelihood estimates. The MCWM algorithm has better mixing properties, but tends to give conservative approximations of the posterior and is still expensive. A new method is developed to accelerate the GIMH method by using a Gaussian process (GP) approximation to the log-likelihood and train this GP using a short pilot run of the MCWM algorithm. This new method called GP-GIMH is illustrated on simulated data from a stochastic volatility and a gene network model. The new approach produces reasonable posterior approximations in these examples with at least an order of magnitude improvement in computing time. Code to implement the method for the gene network example can be found at http://www.runmycode.org/companion/view/2663.

Recent Bayesian approaches for spatial analysis of 2-D images with application to environmental modelling

From remote sensing of the environment, to brain scans in medicine, the growth in the use of image data has motivated a parallel increase in statistical techniques for analysing these images. A particular area of growth has been in Bayesian models and corresponding computational methods. Bayesian approaches have been proposed to address the gamut of supervised and unsupervised inferential aims in image analysis. In this article we provide a general review of these approaches, with a focus on unsupervised analysis of 2-D images. Four exemplar methods that canvas the broad aims of image modelling and analysis are described. An exposition of these approaches is provided by applying them to an environmental case study involving the use of satellite data to assess water quality in the Great Barrier Reef, Australia. The techniques considered in detail are hidden Markov random fields (MRF), Gaussian MRF, Poisson/gamma random fields, and Voronoi tessellations. We also consider a variety of enabling computational algorithms, including MCMC, variational Bayes and integrated nested Laplace approximations. We compare the different aims and inferential capabilities of the models and discuss the advantages and drawbacks of the corresponding computational algorithms.

An external field prior for the hidden Potts model with application to cone-beam computed tomography

In images with low contrast-to-noise ratio (CNR), the information gain from the observed pixel values can be insufficient to distinguish foreground objects. A Bayesian approach to this problem is to incorporate prior information about the objects into a statistical model. A method for representing spatial prior information as an external field in a hidden Potts model is introduced. This prior distribution over the latent pixel labels is a mixture of Gaussian fields, centred on the positions of the objects at a previous point in time. It is particularly applicable in longitudinal imaging studies, where the manual segmentation of one image can be used as a prior for automatic segmentation of subsequent images. The method is demonstrated by application to cone-beam computed tomography (CT), an imaging modality that exhibits distortions in pixel values due to X-ray scatter. The external field prior results in a substantial improvement in segmentation accuracy, reducing the mean pixel misclassification rate for an electron density phantom from 87% to 6%. The method is also applied to radiotherapy patient data, demonstrating how to derive the external field prior in a clinical context. External field prior improves image segmentation accuracy.Manual segmentation of one image is used as a prior for subsequent images.Applicable to longitudinal imaging, such as image-guided radiation therapy.

Constructing a clinical decision-making framework for image-guided radiotherapy using a Bayesian Network

A decision-making framework for image-guided radiotherapy (IGRT) is being developed using a Bayesian Network (BN) to graphically describe, and probabilistically quantify, the many interacting factors that are involved in this complex clinical process. Outputs of the BN will provide decision-support for radiation therapists to assist them to make correct inferences relating to the likelihood of treatment delivery accuracy for a given image-guided set-up correction. The framework is being developed as a dynamic object-oriented BN, allowing for complex modelling with specific subregions, as well as representation of the sequential decision-making and belief updating associated with IGRT. A prototype graphic structure for the BN was developed by analysing IGRT practices at a local radiotherapy department and incorporating results obtained from a literature review. Clinical stakeholders reviewed the BN to validate its structure. The BN consists of a sub-network for evaluating the accuracy of IGRT practices and technology. The directed acyclic graph (DAG) contains nodes and directional arcs representing the causal relationship between the many interacting factors such as tumour site and its associated critical organs, technology and technique, and inter-user variability. The BN was extended to support on-line and off-line decision-making with respect to treatment plan compliance. Following conceptualisation of the framework, the BN will be quantified. It is anticipated that the finalised decision-making framework will provide a foundation to develop better decision-support strategies and automated correction algorithms for IGRT.

Automated replication of cone beam CT‐guided treatments in the Pinnacle3 treatment planning system for adaptive radiotherapy

Time‐consuming manual methods have been required to register cone‐beam computed tomography (CBCT) images with plans in the Pinnacle3 treatment planning system in order to replicate delivered treatments for adaptive radiotherapy. These methods rely on fiducial marker (FM) placement during CBCT acquisition or the image mid‐point to localise the image isocentre. A quality assurance study was conducted to validate an automated CBCT‐plan registration method utilising the Digital Imaging and Communications in Medicine (DICOM) Structure Set (RS) and Spatial Registration (RE) files created during online image‐guided radiotherapy (IGRT).

Bayesian computational methods for spatial analysis of images

This thesis introduces a new way of using prior information in a spatial model and develops scalable algorithms for fitting this model to large imaging datasets. These methods are employed for image-guided radiation therapy and satellite based classification of land use and water quality. This study has utilized a pre-computation step to achieve a hundredfold improvement in the elapsed runtime for model fitting. This makes it much more feasible to apply these models to real-world problems, and enables full Bayesian inference for images with a million or more pixels.