Quantifying the Uncertainty of Contour Maps

This paper proposes an asymptotic embedding method for the dynamic reconstruction of spatially varying fields. By assuming that the spatial field is the solution to an elliptic partial differential equation, then the elliptic PDE is embedded into a parabolic PDE which represents the time-varying estimator. An important advantage of the dynamic estimation scheme is the significant reduction in the use of sensing devices needed to reconstruct the spatial field. Static estimation schemes impose stringent conditions on the regularity of a regression matrix, which links the basis functions to the number of measurements. To further improve the performance of the dynamic estimator, a guidance scheme is proposed that repositions mobile sensors within the spatial field, which is linked to the performance of the dynamic estimator. Extensions to collaborative estimation and optimization of the placement of static sensors are also summarized to provide an integrated account on all facets of optimal dynamic estimation of spatial fields. Numerical simulations for spatial fields in one and two spatial dimensions are included along with a comparison of static reconstruction as quantified by the number of sensing devices required and the relative error.

Evaluating the power densities emitted by GSM1800 and GSM900 BTS antennas isconducted via two methods. Measurements are carried out in half a square meter grids around twoantennas. CST Microwave STUDIO software is employed to estimate the power densities in order fordetailed antenna and tower modeling and simulation of power density. Finally, measurements obtainedfrom computational and experimental methods were compared through the contour lines using thestatistical Surfer software. After measuring and simulating all values, it turns out that power density isgenerally lower than the permissible exposure limits although exceeds the limits in some sample points. According to the measurements, simulation error in stations GSM900 and GSM1800 are 10% and 8%,respectively. Findings from contour-line-maps illustrates that direct measurement method follows thesame emission pattern as the computational method does. It validates the computational approach andthe models attained for BTS power density estimation.

Medical image segmentation is the basis of medical image three-dimensional reconstruction. The accuracy of image segmentation directly affects the results of image 3D reconstruction. According to the grayscale feature of different tissues in CT image and the image symmetry, a new algorithm of obtaining tissue contour lines is put forward. Using edge detection with specific threshold and improved contour tracking algorithm to segment image, then obtain the contour lines. Experimental results show this novel algorithm can improve the efficiency and the precision of contour lines.

A common interaction with medical 2D and 3D images is to apply windows, icons, menus and pointers (WIMP) paradigm; however, WIMP might not be convenient in some situation, such as in the operating room where sterilization becomes an issue. Thus, in this project, the controller-free natural interaction on 3D medical image viewer software through a Kinect camera will be developed to replace WIMP such that surgeons can view medical images in the operating room without contacting mouse and keyboard. The interaction will include both body and figure gestures. The 3D medical image viewer software will be based on the Visualization ToolKit (VTK) library to visualize 3D image data in the DICOM (Digital Imaging and Communications in Medicine) format. Moreover, suitable controller patterns will be designed according to practical use in the operating room.

Sampling of physical fields with mobile sensor is an emerging area. In this context, this work introduces and proposes solutions to a fundamental question: can a spatial field be estimated from samples taken at unknown sampling locations? Unknown sampling location, sample quantization, unknown bandwidth of the field, and presence of measurement-noise present difficulties in the process of field estimation. In this work, except for quantization, the other three issues will be tackled together in a mobile-sampling framework. Spatially bandlimited fields are considered. It is assumed that measurement-noise affected field samples are collected on spatial locations obtained from an unknown renewal process. That is, the samples are obtained on locations obtained from a renewal process, but the sampling locations and the renewal process distribution are unknown. In this unknown sampling location setup, it is shown that the mean-squared error in field estimation decreases as $O(1/n)$ where $n$ is the average number of samples collected by the mobile sensor. The average number of samples collected is determined by the inter-sample spacing distribution in the renewal process. An algorithm to ascertain spatial field's bandwidth is detailed, which works with high probability as the average number of samples $n$ increases. This algorithm works in the same setup, i.e., in the presence of measurement-noise and unknown sampling locations.

We present a fully decentralized algorithm that is inspired by sparse Bayesian learning (SBL) and can be used for non-parametric sparse estimation of unknown spatial functions - spatial fields - with wireless sensor networks (WSNs). It is assumed that a spatial field is represented as a linear combination of weighted fixed basis functions. By exploiting the similarity between the topology of a WSN and the proposed probabilistic graphical model for distributed SBL, a combination of variational inference and loopy belief propagation (LBP) is used to obtain the weights and the sparse subset of relevant basis functions. The algorithm requires only transmission between neighboring sensors and no multi-hop communication is needed. Furthermore, it does not rely on a fixed network structure and no information about the total number of sensors in the network is necessary. Due to consensus in the weight parameters between neighboring sensors, it is demonstrated that also the sparsity patterns of relevant basis functions generally agree. The effectiveness of the proposed algorithm is demonstrated with synthetic data.

This paper anticipate a deep learning (DL)-located framework for medical image captioning (IC) utilizing a transformerlocated model. The system is anticipated like it integrates two important parts, a Vision Transformer (VT) for processing ocular data and fetching features from images, and a decoder based on Generative Pre-trained Transformer 2 (GPT-2), that transform these characteristics into understandable, contextually significant natural language depictions. These two works composed to aid automatic creation of textual reports matching medical images. To ensure output quality and relevance, training is conducted on a carefully selected dataset that includes matched medical images and clinical descriptions. The model's performance was evaluated in this work using the IU X-Ray dataset, a publicly accessible and extensively used benchmark in the medical imaging community. Preliminary diagnostic reporting should be generated automatically, which will reduce labor costs, increase workflow effectiveness, shorten reporting times, and improve consistency with medical records. The transformer-based approach manages long-range dependencies in both the picture and text domains, in contrast to conventional CNN–RNN designs.Better diagnostic interpretability results from more context-aware and semantically rich medical image captions.In this research, we offer an end-to-end medical picture captioning model that is totally transformer-based and does away with recurrent components like LSTM.The paper's main goal is to make diagnostic reporting easier for radiologists and other medical practitioners. Based on popular evaluation measures including BLEU, METEOR, and ROUGE, the model showed competitive results (98.42% accuracy), suggesting significant potential for practical use.

The prediction of breeding values and phenotypes is of central importance for both livestock and crop breeding. In this study, we analyze the use of artificial neural networks (ANN) and, in particular, local convolutional neural networks (LCNN) for genomic prediction, as a region-specific filter corresponds much better with our prior genetic knowledge on the genetic architecture of traits than traditional convolutional neural networks. Model performances are evaluated on a simulated maize data panel (n = 10,000; p = 34,595) and real Arabidopsis data (n = 2,039; p = 180,000) for a variety of traits based on their predictive ability. The baseline LCNN, containing one local convolutional layer (kernel size: 10) and two fully connected layers with 64 nodes each, is outperforming commonly proposed ANNs (multi layer perceptrons and convolutional neural networks) for basically all considered traits. For traits with high heritability and large training population as present in the simulated data, LCNN are even outperforming state-of-the-art methods like genomic best linear unbiased prediction (GBLUP), Bayesian models and extended GBLUP, indicated by an increase in predictive ability of up to 24%. However, for small training populations, these state-of-the-art methods outperform all considered ANNs. Nevertheless, the LCNN still outperforms all other considered ANNs by around 10%. Minor improvements to the tested baseline network architecture of the LCNN were obtained by increasing the kernel size and of reducing the stride, whereas the number of subsequent fully connected layers and their node sizes had neglectable impact. Although gains in predictive ability were obtained for large scale data sets by using LCNNs, the practical use of ANNs comes with additional problems, such as the need of genotyping all considered individuals, the lack of estimation of heritability and reliability. Furthermore, breeding values are additive by design, whereas ANN-based estimates are not. However, ANNs also comes with new opportunities, as networks can easily be extended to account for additional inputs (omics, weather etc.) and outputs (multi-trait models), and computing time increases linearly with the number of individuals. With advances in high-throughput phenotyping and cheaper genotyping, ANNs can become a valid alternative for genomic prediction.

Medical imaging is becoming a vital component of war on cancer. Tremendous amounts of medical image data are captured and recorded in a digital format during cancer care and cancer research. Facing such an unprecedented volume of image data with heterogeneous image modalities, it is necessary to develop effective and efficient content-based medical image retrieval systems for cancer clinical practice and research. While substantial progress has been made in different areas of content-based image retrieval (CBIR) research, direct applications of existing CBIR techniques to the medical images produced unsatisfactory results, because of the unique characteristics of medical images. In this paper, we develop a new multimodal medical image retrieval approach based on the recent advances in the statistical graphic model and deep learning. Specifically, we first investigate a new extended probabilistic Latent Semantic Analysis model to integrate the visual and textual information from medical images to bridge the semantic gap. We then develop a new deep Boltzmann machine-based multimodal learning model to learn the joint density model from multimodal information in order to derive the missing modality. Experimental results with large volume of real-world medical images have shown that our new approach is a promising solution for the next-generation medical imaging indexing and retrieval system.

The main goal of current engine development is to increase power density and efficiency and to minimize engine emissions. The idea is to obtain the desired power output with a highly charged combustion engine in combination with exhaust gas turbocharging and a very small engine displacement, which is known as downsizing. The selection of a turbocharger is based on the maps of the turbine and compressor, which are usually measured on a test bench. They also provide important boundary conditions on the engine process simulation of a supercharged engine with this turbocharger. In general, a very accurate measurement of the characteristic maps is desired to ensure the best possible matching. However, random and systematic errors have an impact on the measurement results. In order to assess the quality of the measured and calculated values, it is necessary to determine the uncertainties of the measurement variables as accurately as possible; particularly, the error propagation in calculating the efficiencies. The uncertainties are based on a systematic uncertainty component of the sensor and the confidence interval. In this way, the measurement uncertainty is estimated by linear and geometric combination of the calculated random and systematic uncertainties. After that, the respective uncertainty contributions and the identification of the relevant parameters that influence the resulting measurement uncertainty are evaluated. Knowing the measurement uncertainties of the characteristic maps of a turbocharger, the influence on engine operation will be determined with a one-dimensional engine process simulation model. Consequently, the determined measurement uncertainty will be applied as a deviation on the efficiencies and will be investigated in a GT POWER simulation. The impact of the measurement uncertainty on the engine performance is shown on the basis of load steps.

The main goal of current engine development is to increase power density and efficiency and to minimize engine emissions. The idea is to obtain the desired power output with a highly charged combustion engine in combination with exhaust gas turbocharging and a very small engine displacement, so-called downsizing. The selection of a turbocharger is based on the maps of the turbine and compressor, which are usually measured on a test bench. They also provide important boundary conditions on the engine process simulation of a supercharged engine with this turbocharger. In general, a very accurate measurement of the characteristic maps is desired to ensure the best possible matching. However, random and systematic errors have an impact on the measurement results. In order to assess the quality of the measured and calculated values, it is necessary to determine the uncertainties of the measurement variables as accurately as possible, particularly the error propagation in calculating the efficiencies. The uncertainties are based on a systematic uncertainty component of the sensor and the confidence interval. In this way the measurement uncertainty is estimated by linear and geometric combination of the calculated random and systematic uncertainties. After that, the respective uncertainty contributions and the identification of the relevant parameters that influence the resulting measurement uncertainty are evaluated. Knowing the measurement uncertainties of the characteristic maps of a turbocharger, the influence on engine operation will be determined with a one-dimensional engine process simulation model. Consequently, the determined measurement uncertainty will be applied as a deviation on the efficiencies and will be investigated in a GT POWER simulation. The impact of the measurement uncertainty on the engine performance is shown on the basis of load steps.

Spatial field sampling with mobile sensor has recently been addressed in the literature. This work introduces and proposes solution to a fundamental question: can a spatial field be estimated from samples taken by a mobile sensor at unknown sampling locations along a path? Spatially one-dimensional and bandlimited, and temporally fixed fields are considered. It is assumed that field samples are collected on spatial locations realized by an unknown renewal process. That is, the sampling locations and the inter-sample distribution in the renewal process are both unknown. It is shown that average mean-squared error in field estimation decreases as O(1/n) where n is the sampling-rate employed by the mobile sensor. The sampling rate can be increased by controlling the mean value of the inter-sample spacing by a location unaware mobile sensor.

Estimate of coronal magnetic fields from contour maps of the intensity and polarization degree of the burst on 1992 October 27

In this thesis, we focus on two important medical image processing and analysis problems: multimodal image registration and cerebrovascular segmentation. First, for the purpose of improving the robustness of image registration, we propose two novel registration objective functions and a new technique for better optimization performance. The first novel objective function is a combination of the multi-dimensional mutual information and an angle measure on the maximum distance-gradient (MDG) vector field. The MDG vector field is a new spatial feature field designed for registration tasks. It encodes both local edge information and globally defined spatial information related to the intensity difference, the distance, and the direction of a voxel to a certain object boundary point. This objective function then integrates both the magnitude and the orientation information of the MDG feature into the image registration process. The other novel objective function can be obtained if the statistical joint intensity mappings estimated from the pre-aligned training image pairs are available. This a priori knowledge can then be used as reference to guide the multi-resolution based registration process, by minimizing a new Kullback-Leibler distance (KLD) based similarity measure between the observed joint intensity distribution and the reference joint intensity distribution. To achieve better optimization, which evidently affects the robustness of a registration process, we propose a general technique based on random transformation disturbance to assist optimization procedures to avoid local optima, and hence to achieve robust optimization results. The technique is based on different properties of the local neighborhood of a global optimum and those of local optima. Second, this thesis presents a statistical segmentation technique for the extraction of vasculatures in three-dimensional rotational angiography (3D-RA). This method uses maximum intensity projections (MIP) to improve the accuracy and robustness of threshold estimation using the expectation maximization (EM) algorithm. It is fully automatic and computationally efficient. Finally, encouraging experimental results of the proposed approaches on both synthetic and clinical medical images are presented in the thesis.

Deep learning (DL) models are effective in leveraging latent representations from MR data, emerging as state-of-the-art solutions for accelerated MRI reconstruction. However, challenges arise due to the inherent uncertainties associated with undersampling in k-space, coupled with the over- or under-parameterized and opaque nature of DL models. Addressing uncertainty has thus become a critical issue in DL MRI reconstruction. Monte Carlo (MC) inference techniques are commonly employed to estimate uncertainty, involving multiple reconstructions of the same scan to compute variance as a measure of uncertainty. Nevertheless, these methods entail significant computational expenses, requiring multiple inferences through the DL model. In this context, we propose a novel approach to uncertainty estimation during MRI reconstruction using a pixel classification framework. Our method, PixCUE (Pixel Classification Uncertainty Estimation), generates both the reconstructed image and an uncertainty map in a single forward pass through the DL model. We validate the efficacy of this approach by demonstrating that PixCUE-generated uncertainty maps exhibit a strong correlation with reconstruction errors across various MR imaging sequences and under diverse adversarial conditions. We present an empirical relationship between uncertainty estimations using PixCUE and established reconstruction metrics such as NMSE, PSNR, and SSIM. Furthermore, we establish a correlation between the estimated uncertainties from PixCUE and the conventional MC method. Our findings affirm that PixCUE reliably estimates uncertainty in MRI reconstruction with minimal additional computational cost.