Title: Dose Optimization for Pediatric Renal SPECT Imaging
Abstract: Like any real-world problem, the design of an imaging system always requires tradeoffs. For medical imaging modalities using ionization radiation, a major tradeoff is between diagnostic image quality (IQ) and risk to the patient from absorbed radiation dose. In nuclear medicine, reducing the radiation dose to the patient will always result in increased Poisson noise in the image. At the same time, reducing the radiation dose (RD), below some level at least, will always result in reduced risk of adverse effects to the patient. The overall goal of this research is to propose a rigorous IQ-RD tradeoff analysis method for pediatric nuclear medicine renal imaging. However, the methodologies developed in this proposal can also be applied to other nuclear medicine imaging applications and other important medical modalities involving ionization radiation such as computed tomography and planar X-rays.
Balancing the tradeoffs between RD and IQ is especially important for children, as they are often considered more vulnerable to radiation than adults. In nuclear medicine imaging, reducing the RD requires reducing the administered activity (AA). Lower AA results in increased Poisson noise in the images or requires longer acquisition durations to maintain the noise level. In pediatric nuclear medicine, it is desirable to use the lowest AA and the shortest acquisition duration that gives sufficient IQ for clinical diagnosis. In current clinical practice, AA for pediatric molecular imaging is often based on the North American consensus guidelines (U.S.) and the European pediatric dosage card (Europe). Both of these dosing guidelines involve scaling the adult AA by patient weight subject to upper and lower constraints on the AA. However, these guidelines were developed based on expert consensus or rough estimates (estimated count rates) of IQ rather than rigorous, objective measures of performance on the diagnostic task.
In this research, we propose a general framework for optimizing RD with task-based assessment of IQ. Here, IQ is defined as an objective measure of the user performing the diagnostic task that the images were acquired to answer. Specifically, we propose to establish relationships between AA, acquisition duration, measures of body habitus, and IQ for pediatric patients undergoing renal molecular imaging procedures. To investigate IQ as a function of renal defect detectability, we have developed a projection image database modeling imaging of 99mTc-DMSA, a renal function agent. The database uses a highly-realistic population of pediatric phantoms with anatomical and body morphological variations. Using the developed projection image database, we have explored patient factors that affect IQ and are currently in the process of determining relationships between IQ and AA (IQ-AA curve) in terms of these found factors. Our preliminary data have shown that the current weight-based guidelines, based on scaling the AA by patient weight, are not optimal in the sense that they do not give the same image quality for patients with the same weight. Furthermore, we have found that factors that are more local to the target organ may be more robust than weight for estimating the AA needed to provide a constant IQ across a population of patients. In the case of renal imaging, we have discovered that girth is more robust than weight in predicting AA needed to provide a desired IQ. In addition, in order to simulate a full clinical multi-slice detection task (just like what a nuclear medicine physician would do), we propose to develop a CNN-based model observer. We will perform human observer studies to verify and calibrate the developed model observers used to generate the IQ-AA curves. The results of this proposal will provide the data needed by standards bodies to develop improved dosing guidelines for pediatric molecular imaging that result in more consistent image quality and absorbed dose.
Title: Modeling Cellular Events: Chemotaxis and Aneuploidy
Abstract: Biology is the ‘study of complex natural things’ and the biologists are mostly interested in details of those complexity in a system. But often a simpler mathematical model is proved to be very efficient in deciphering the underlying basic working principle of the system. Despite the usefulness, these models are often criticized for not being able to explain the sufficient details of the wide range of experimental observations of different cases of pharmacological/genetic perturbations.
Title: Extending the potential of thin-film optoelectronics via optical and photonic engineering
Project summary: Thin-film optoelectronics using solution-processed materials have become a strong research focus in recent decades. These technologies have demonstrated convenience and versatility, due to their solution-processed nature, in a wide range of applications such as solar power harvesting, photodetection, light emitting devices and even lasing. Some of the variants of these materials also enabled and dominate the field of flexible electronics, especially for display technologies, achieving large-scale industrialization and commercialization years ago specifically in applications where their conventional counterparts – bulk semiconductors – are limited. The development of optoelectronics applications using organic materials, colloidal quantum dots, perovskites, etc., has been made possible by research progress in materials and chemical engineering of the active material itself, as well as in optical and photonic engineering in the device architecture and related structures. The focus of this project is mainly on the latter set of approaches applied to lead chalcogenide-based colloidal quantum dot thin films.
Colloidal quantum dots (CQDs) are a type of semiconductor material in the form of nanocrystals (1-10 nm in diameter) of the corresponding bulk material. The spatial confinement of electrons and holes leads to significantly reconstructed energy band structures. Usually this manifests as a series of discrete energy levels above or below the corresponding bulk conduction and valence band edges, instead of the corresponding semi-continuum of states observed in bulk semiconductors. The spacings between the discrete energy levels are highly dependent on the size of the quantum dots, which at the same time determines the properties of optical transitions responsible for absorption (Figure 1b), modulation of the refractive index, etc. In this sense, CQDs are considered “tunable” by controlling the ensemble so that it predominantly consisting CQDs of one desired shape and size.
CQDs are solution-processed materials. The processing of CQDs starts from synthesis using solutions containing metal-organic precursors. The controlled growth of nanocrystals results in a dispersion of pristine CQDs in certain solvents. After that, the CQDs are purified and chemically treated to modify their surface ligands, through a series of precipitation, redispersion, phase transfer and concentration steps. The deposition of films of CQDs onto desired substrates is achieved by solution-compatible techniques such as spin-casting, blade coating and screen printing. A functional CQD film is usually 10-500 nm thick depending on its application and is usually preceded and/or succeeded by the deposition of other electronically functional device layers.
Lead sulfide (PbS) CQDs are widely used for applications involving solar photon absorption and resulting energy conversion. In the example of a CQD solar cell, PbS CQDs with effective band gaps of 1.3 eV are chosen as the active material. The full device utilizes a p-n or p-i-n structure, and a typical device architecture consists of a transparent conductive oxide (TCO) electrode layer, an electron transport layer (ETL), the absorbing PbS CQD film, a hole transport layer (HTL) and metal top electrode. Similar structures are also used in photodetectors and light emitting diodes, with critical layers substituted.
For the first section of the project, we studied and exploited the color reproduction capabilities using reflective interference from CQD solar cells, while maintaining high photon absorption and current generation. The second section is aimed at exploring the possibility of simultaneously controlling the spectral reflection, transmission and absorption of thin film optoelectronics using embedded photonic crystal structures in CQD films and other highly absorptive materials. In the third section, we devised and built a 2D multi-modal scanning characterization system for spatial mapping of photoluminescence (PL), transient photocurrent and transient photovoltage from a realistically large device area with micron-resolution. The last section of the project focuses on economical and scalable solar concentration solutions for CQD and other thin film solar cells.
We mostly limit our discussion and demonstration to PbS CQD solar cells within the
scope of this proposal; however, it is worth pointing out that the techniques and
principles described below could be applied to most optoelectronic materials that share
the solution-compatible deposition and processing procedures.
Title: New Diagnostic and Therapeutic Tools for Intravascular Magnetic Resonance Imaging (IVMRI)
Abstract: Intravascular (IV) magnetic resonance imaging (IVMRI) is a developing technology that uses minimally-invasive MRI coils to guide diagnosis and treatment. The combination of signal-to-noise (SNR) enhancement from the microscopic MRI local coils and the multi-contrast mechanisms provided by MRI has enlarged the possibilities of high-resolution imaging-guided diagnosis and treatment of atherosclerosis and nearby or surrounding cancers. Recent years have seen the development of many advanced MRI techniques including MRI thermometry and real-time MRI, yet the development of procedures that apply these advances to intravascular MRI remain challenging.
Among interventional diagnostic techniques, MRI endoscopy is an IVMRI technique that transfers MRI from the laboratory frame-of-reference to the IV-coil’s frame-of-reference. This enables high-resolution MRI of blood vessels with endoscopic-style functionality. Prior MRI endoscopy work was limited to ~2 frames-per-second (fps), which is not real-time and potentially limiting in clinical applications. Improving the speed of MRI endoscopy further without excessive undersampling artifacts could enable the rapid deployment and advancement of an IVMRI endoscope entirely by MRI guidance to evaluate local, advanced, intra- and extra-vascular disease at high resolution using MRI’s unique multi-contrast and multi-functional assessment capabilities. Furthermore, with its unique capability in high-resolution thermometry, IVMRI is suitable to guide and monitor ablation therapy delivery in disease such as vessel-involving cancers. Prior work using an IVMRI loopless antenna for both MRI and radiofrequency ablation (RFA) was limited in precision and ablated only the tissue in direct contact with the probe. Thus, one goal is to extend IVMRI methods using state-of-the-art real-time MRI acceleration methods to provide MRI endoscopy at a speed comparable to that of existing catherization and optical endoscopy procedures.
A second goal is to provide a minimally-invasive, IV-accessed ablation technology that could provide precision localization and perivascular ablation to render resectable, an otherwise inaccessible or non-resectable cancer with vascular involvement.
To these ends, a Max-Planck Institute (MPI) real-time MRI system employing graphic processing units (GPU) is first adapted to facilitate MRI endoscopy at 10 fps endoscopy with real-time display and is demonstrated in vitro and in vivo. To further improve image quality, we propose to use a neural network (CNN) trained on artifact patterns generated from motionless endoscopy to ameliorate artifacts during real-time imaging. A new method based on generative models and manifold learning is then proposed to optimize image contrast responsive to the varying endoscopic surroundings.
To address the second goal, an intravascular ultrasound ablation transducer is integrated with IVMRI to provide a tool that can also deliver therapy. By integrating an IV high-intensity ultrasound (HIFU) ablation component, the precision and depth of ablation is extended and contact injuries can be avoided. Procedures are developed to evaluate accuracy using ex vivo samples and feasibility is demonstrated in animals in vivo.
Title: Soroban: A Mixed-Signal Neuromorphic Processing in Memory Architecture
Abstract: To meet the scientific demand for future data-intensive processing for every day mundane tasks such as searching via images to the uttermost serious health care disease diagnosis in personalized medicine, we urgently need a new cloud computing paradigm and energy efficient i.e. “green” technologies. We believe that a brain-inspired approach that employs unconventional processing offers an alternative paradigm for BIGDATA computing.
My research aims to go beyond the state of the art processor in memory architectures. In the realm of un-conventional processors, charge based computing has been an attractive solution since it’s introduction with charged-coupled device (CCD) imagers in the seventies. Such architectures have been modified to compute-in-memory arrays that have been used for signal processing, neural networks and pattern recognition using the same underlying physics. Other work has utilized the same concept in the charge-injection devices (CIDs), which have also been used for similar pattern recognition tasks. However, these computing elements have not been integrated with the support infrastructure for high speed input/output commensurate with BIGDATA processing streaming applications. In this work, the CID concept is taken to a smaller CMOS 55nm node and has shown promising preliminary results as a multilevel input computing element for hardware inference applications. A mixed signal charge-based vector-vector multiplier (VMM) is explored which computes directly on a common readout line of a dynamic random-access memory (DRAM). Low power consumption and high area density is achieved by storing local parameters in a DRAM computing crossbar.
Title: Advanced Image Reconstruction and Analysis for Fluorescence Molecular Tomography (FMT) and Positron Emission Tomography (PET)
Abstract: Molecular imaging provides efficient ways to monitor different biological processes noninvasively, and high-quality imaging is necessary in order to fully explore the value of molecular imaging. To this end, advanced image generation algorithms are able to significantly improve image quality and quantitative performance. In this research proposal, we focus on two imaging modalities, fluorescence molecular tomography (FMT) and positron emission tomography (PET), that fall in the category of molecular imaging. Specifically, we studied the following two problems: i) reconstruction problem in FMT and ii) partial volume correction in brain PET imaging.
Reconstruction in FMT: FMT is an optical imaging modality that uses diffuse light for imaging. Reconstruction problem for FMT is highly ill-posed due to photon scattering in biological tissue, and thus, regularization techniques tend to be used to alleviate the ill-posed nature of the problem. Conventional reconstruction algorithms cause oversmoothing which reduces resolution of the reconstructed images. Moreover, a Gaussian model is commonly chosen as the noise model although most FMT systems based on charged-couple device (CCD) or photon multiplier tube (PMT) are contaminated by Poisson noise. In our work, we propose a reconstruction algorithm for FMT using sparsity-initialized maximum-likelihood expectation maximization (MLEM). The algorithm preserves edges by exploiting sparsity, as well as taking Poisson noise into consideration. Through simulation experiments, we compare the proposed method with pure sparse reconstruction method and MLEM with uniform initialization. We show the proposed method holds several advantages compared to the other two methods.
Partial volume correction of brain PET imaging: The so-called partial volume effect (PVE) is caused by the limited resolution of PET systems, reducing quantitative accuracy of PET imaging. Based on the stage of implementation, partial volume correction (PVC) algorithms could be categorized into reconstruction-based and post-reconstruction methods.Post reconstruction PVC methods can be directly implemented on reconstructed PET images and do not require access to raw data or reconstruction algorithms of PET scanners. Many of these methods use anatomical information from MRI to further improve their performance. However, conventional MR guided post-reconstruction PVC methods require segmentation of MR images and assume uniform activity distribution within each segmented region. In this proposal, we develop post-reconstruction PVC method based on deconvolution via parallel level set regularization. The method is implemented with non-smooth optimization based on the split Bregman method. The proposed method incorporates MRI information without requiring segmentation or making any assumption on activity distribution. Simulation experiments are conducted to compare the proposed method with several other segmentationfree method, as well as conventional segmentation-based PVC method. The results show the proposed method outperforms other segmentation-free method and shows stronger resistance to MR information mismatch compared to conventional segmentation-based PVC method.
Title: Statistical Modeling and analysis of allele-specific DNA methylation at the haplotype level
Abstract: Epigenetics is the branch of biology concerned with the study of phenotypical changes due to alterations of DNA, maintained during cell division, excluding modifications of the sequence itself. Epigenetic information includes DNA methylation, histone modifications, and higher order chromatin structure among others. DNA methylation is a stable epigenetic mechanism that chemically marks the DNA by adding methyl groups at individual cytosines immediately adjacent to guanines (CpG sites). Methylation marks are used to identify cell-type specific aspects of gene regulation, since marks located within a gene promoter or enhancer typically act to repress gene transcription, whereas promoter or enhancer demethylation is associated with gene activation. Notably, patterns of methylation marks are highly polymorphic and stochastic, containing information about a broad range of normal and aberrant biological processes, such as development and differentiation, aging, and carcinogenesis.
The epigenetic information content of two homologous chromosomal regions need not be the same. For example, it is well established that the ability of a cell to methylate the promoter region of a specific copy of a gene (an allele), is crucial for proper development. In fact, many known phenotypical traits stem from allele-specific epigenetic marks. Moreover, some allele-specific epigenetic differences have been found to be associated with local genetic differences between copies of a chromosome. Thus, developing a framework for studying such epigenetic differences in diploid organisms is our main goal. More specifically, our objective is to develop a statistical method that can be used to detect regions in the genome, with genetic differences between homologous chromosomes, in which there are biologically relevant differences in DNA methylation between alleles.
State of the art methods for allele-specific methylation modeling and analysis have critical shortcomings rendering them unsuitable for this type of analysis. We present a statistical physics inspired model for allele-specific methylation analysis that contains a sensible number of parameters, considering the limited sample size in whole genome bisulfite sequencing data, which is rich enough to capture the complexity in the data. We demonstrate the appropriateness of this model for allele-specific methylation analysis using simulation data as well as real data. Using our model, we compute mean methylation level differences between alleles, as well as information-theoretic quantities, such as the entropy of the methylation state in each allele and the mutual information between the methylation state and the allele of origin, and assess the statistical significance of each quantity by learning the null distribution from the data. This complementary set of statistics allows for an unparalleled level of insight in subsequent biological analysis. As a result, the developed framework provides an unprecedented descriptive power to characterize (i) the circumstances under which allele-specific methylation events arise, and (ii) the cis-effect, or lack of thereof, that genetic mutations have on DNA methylation.
Title: A Practical and Efficient Multi-Stream Framework for End-to-End Speech Recognition
Abstract: The multi-stream paradigm in Automatic Speech Recognition (ASR) considers scenarios where parallel streams carry diverse or complementary task-related knowledge. In these cases, an appropriate strategy to fuse streams or select the most informative source is necessary. In recent years, with the increasing use of Deep Neural Networks (DNNs) in ASR, End-to-End (E2E) approaches, which directly transcribe human speech into text, have received greater attention. In this proposal, a multi-stream framework is present based on joint CTC/Attention E2E model, where parallel streams are represented by separate encoders aiming to capture diverse information. On top of the regular attention networks, a secondary stream-fusion network is introduced to steer the decoder toward the most informative encoders.
Two representative framework have been proposed, which are MultiEncoder Multi-Resolution (MEM-Res) and Multi-Encoder Multi-Array (MEM-Array), respectively. Moreover, with an increasing number of streams (encoders) requiring substantial memory and massive amounts of parallel data, a practical two-stage training scheme is further proposed in this work. Experiments are conducted on various corpora including Wall Street Journal (WSJ), CHiME-4, DIRHA and AMI. Compared with the best single-stream performance, the proposed framework has achieved substantial improvement, which also outperforms various conventional fusion strategies.
The future plan aims to improve robustness of the proposed multistream framework. Measuring performance of an ASR system without ground-truth could be beneficial in multi-stream scenarios to emphasize on more informative streams than corrupted ones. In this proposal, four different Performance Monitoring (PM) techniques are investigated. The preliminary results suggest that PM measures on attention distributions and decoder posteriors are well-correlated with true performances. Integration of PM measures and more sophisticated fusion mechanism in multi-stream framework will be the focus for future exploration.
Title: Deep Learning-based Novelty Detection
Abstract: In recent years, intelligent systems powered by artificial intelligence and computer vision that perform visual recognition have gained much attention. These systems observe instances and labels of known object classes during training and learn association patterns that can be
used during inference. A practical visual recognition system should first determine whether an observed instance is from a known class. If it is from a known class, then the identity of the instance is queried through classification. The former process is commonly known as novelty detection (or novel class detection) in the literature. Given a set of image instances from known classes, the goal of novelty detection is to determine whether an observed image during inference belongs to one of the known classes.
We consider one-class novelty detection, where all training data are assumed to belong to a single class without any finer-annotations available. We identify limitations of conventional approaches in one-class novelty detection and present a Generative Adversarial Network(GAN) based solution. Our solution is based on learning latent representations of in-class examples using a denoising auto-encoder network. The key contribution of our work is our proposal to explicitly constrain the latent space to exclusively represent the given class. In order to accomplish this goal, firstly, we force the latent space to have bounded support by introducing a tanh activation in the encoder’s output layer. Secondly, using a discriminator in the latent space that is trained adversarially, we ensure that encoded representations of in-class examples resemble uniform random samples drawn from the same bounded space. Thirdly, using a second adversarial discriminator in the input space, we ensure all randomly drawn latent samples generate examples that look real.
Finally, we introduce a gradient-descent based sampling technique that explores points in the latent space that generate potential out-of-class examples, which are fed back to the network to further train it to generate in-class examples from those points. The effectiveness of the proposed method is measured across four publicly available datasets using two one-class novelty detection protocols where we achieve state-of-the-art results.
Title: Towards a better understanding of spoken conversations: Assessment of sentiment and emotion
Abstract: In this talk, we present our work on understanding the emotional aspects of spoken conversations. Emotions play a vital role in our daily life as they help us convey information impossible to express verbally to other parties.
While humans can easily perceive emotions, these are notoriously difficult to define and recognize by machines. However, automatically detecting the emotion of a spoken conversation can be useful for a diverse range of applications such as human-machine interaction and conversation analysis. In this work, we considered emotion recognition in two particular scenarios. The first scenario is predicting customer sentiment/satisfaction (CSAT) in a call center conversation, and the second consists of emotion prediction in short utterances.
CSAT is defined as the overall sentiment (positive vs. negative) of the customer about his/her interaction with the agent. In this work, we perform a comprehensive search for adequate acoustic and lexical representations.
For acoustic representation, we propose to use the x-vector model, which is known for its state-of-the-art performance in the speaker recognition task. The motivation behind using x-vectors for CSAT is we observed that emotion information encoded in x-vectors affected speaker recognition performance. For lexical, we introduce a novel method, CSAT Tracker, which computes the overall prediction based on individual segment outcomes. Both methods rely on transfer learning to obtain the best performance. We classified using convolutional neural networks combining the acoustic and lexical features. We evaluated our systems on US English telephone speech from call center data. We found that lexical models perform better than acoustic models and fusion of them provided significant gains. The analysis of errors uncovers that the calls where customers accomplished their goal but were still dissatisfied are the most difficult to predict correctly. Also, we found that the customer’s speech is more emotional compared to the agent’s speech.
For the second scenario of predicting emotion, we present a novel approach based on x-vectors. We show that adapting the x-vector model for emotion recognition provides the best-published results on three public datasets.