This paper introduces a Sliding Mode Observer (SMO) based Finite Control Set Model Predictive Control (FCSMPC) strategy. A digital model of the system, implemented in Matlab/Simulink, is utilized to verify the effecti...
This paper introduces a Sliding Mode Observer (SMO) based Finite Control Set Model Predictive Control (FCSMPC) strategy. A digital model of the system, implemented in Matlab/Simulink, is utilized to verify the effectiveness of the proposed approach. The simulation results, encompassing both steady-state and dynamic analyses, are discussed and benchmarked against traditional FCS-MPC. The results highlight improved total harmonic distortion (THD) values, superior dynamic response, and more stable performance when integrating SMO into the FCS-MPC methodology.
Image-based 3D reconstruction is one of the most important tasks in computer Vision with many solutions proposed over the last few decades. The objective is to extract metric information i.e. the geometry of scene obj...
详细信息
Image-based 3D reconstruction is one of the most important tasks in computer Vision with many solutions proposed over the last few decades. The objective is to extract metric information i.e. the geometry of scene objects directly from images. These can then be used in a wide range of applications such as film, games, virtual reality, etc. Recently, deep learning techniques have been proposed to tackle this problem. They rely on training on vast amounts of data to learn to associate features between images through deep convolutional neural networks and have been shown to outperform traditional procedural techniques. In this paper, we improve on the state-of-the-art two-view structure-from-motion(SfM) approach of [11] by incorporating 4D correlation volume for more accurate feature matching and reconstruction. Furthermore, we extend it to the general multi- view case and evaluate it on the complex benchmark dataset DTU [5]. Quantitative evaluations and comparisons with state-of-the-art multi-view 3D reconstruction methods demonstrate its superiority in terms of the accuracy of reconstructions.
In the field of medical imaging analysis, particularly in interpreting chest X-rays, deep learning models have shown remarkable progress. Nonetheless, these models often face challenges such as limited annotation and ...
详细信息
ISBN:
(数字)9798350386226
ISBN:
(纸本)9798350386233
In the field of medical imaging analysis, particularly in interpreting chest X-rays, deep learning models have shown remarkable progress. Nonetheless, these models often face challenges such as limited annotation and inadequate utilization of public data resources. This is particularly apparent with databases containing multimodal data, such as images and medical reports, where the effective integration of this multimodal information remains difficult. To address these limitations, we propose the Neighbor-Assisted Multimodal Attention Network (NAMAN), a novel approach designed to leverage retrieval augmentation techniques to enhance disease classification performance. NAMAN combines nearest neighbor search with multimodal fusion, utilizing both visual features from similar X-ray images and textual information from corresponding medical records. The experimental results demonstrate the efficacy of incorporating retrieved neighbor information and multimodal integration mechanisms in NAMAN. Our ablation studies offer insights into the optimal configuration of the model, including the effects of various attention mechanisms and the number of retrieved neighbors. This work contributes to the expanding field of retrieval-augmented approaches in medical imaging, presenting a promising avenue for leveraging large-scale, multimodal medical databases to enhance diagnostic accuracy and reliability.
With the rising prevalence of smart homes, there's an increasing demand for comprehensive automation solutions to mitigate fire risks, especially when homeowners are absent or in homes with elderly residents. This...
详细信息
This study examines the decline in the usage of the Javanese language, which has experienced a decrease in the number of speakers from approximately 82 million in 2007 to 68.2 million in 2015. The convergence of the J...
This study examines the decline in the usage of the Javanese language, which has experienced a decrease in the number of speakers from approximately 82 million in 2007 to 68.2 million in 2015. The convergence of the Javanese script and Optical Character Recognition (OCR) technology is proposed as a solution, allowing for the preservation and accessibility of the Javanese script in the digital age. The integration of Convolutional Neural Networks (CNNs) in Javanese script classification achieved a high accuracy rate of 92.95% in identifying positive and negative cases. The dataset used for training consisted of 8440 sample images, which were divided into 20 subfolders for training and testing. The results presented in Table IV demonstrate the successful implementation of the classification model, achieving a 98.87% sensitivity, 100% precision, and 98.88% specificity. This research contributes to the preservation and understanding of the culturally significant Javanese script while addressing the decline in its usage.
Today, medical imaging techniques are widely used to detect a variety of human conditions and diseases. To speed up the diagnostic process, systems are often automated using deep learning methods, which have been prov...
详细信息
ISBN:
(数字)9798350363739
ISBN:
(纸本)9798350363746
Today, medical imaging techniques are widely used to detect a variety of human conditions and diseases. To speed up the diagnostic process, systems are often automated using deep learning methods, which have been proven to yield outstanding results. However, these models may be limited to a specific domain, prone to overfitting, and difficult to update. In this paper, we present a transfer learning approach that achieves high accuracy for detecting anomalies in both MRI and CT images. Our system utilizes the ConvNeXt network, inspired by the Swin Transformer architecture, as a high-level feature extractor from processed and augmented medical images. The resulting data is then optimized using the Edited Nearest Neighbors algorithm, which selects the most effective observations and discards the rest. Finally, the data is passed to the K-Nearest Neighbors classifier to detect anomalies in numerically represented images. Our model achieves high performance for two different types of medical images and can easily be updated without the need for retraining.
Distributed agile development comes with a lot of challenges in particular as it has to do with agile teams working together from different geological locations on the same project. Most probably it happens due to a l...
详细信息
Distributed agile development comes with a lot of challenges in particular as it has to do with agile teams working together from different geological locations on the same project. Most probably it happens due to a lack of visibility in the complex development and deployment process, poor communication, and unavailability of the development team and corresponding customer in the same place. These factors affect the performance of the team and increase the overall cost of development. To mitigate all these aspects, we proposed a cloud computing-based Infrastructure which is a combination of both agile as well as cloud computing technology named ‘CBAI’. The proposed Infrastructure assists the team members to work efficiently even if they are different geo-locations without burdening the cost. It provides the basic structure for global agile development and is also efficient in reducing the technical liability, and the need for project backlog.
Advanced battery management systems (ABMSs) rely on mathematical models to ensure high battery safety and performance. One of the key tasks of a BMS is state estimation. In the following, we consider a single lithium-...
详细信息
Advanced battery management systems (ABMSs) rely on mathematical models to ensure high battery safety and performance. One of the key tasks of a BMS is state estimation. In the following, we consider a single lithium-ion cell described with a dual polarization equivalent circuit model. To consider a realistic scenario, where the parameters have been identified from experimentally collected data, both parametric and measurement uncertainties are taken into account in the model. In particular, unknown but bounded uncertainties are assumed. In this setup, we address state estimation through a set-based approach using Constrained Zonotopes (CZ). Due to the model nonlinearities, a method able to propagate CZ through nonlinear mappings is demanded. Within this context, mean value and first-order Taylor CZ-based extensions were proposed which, however, might lead to conservative overestimation due to the sensitivity to the wrapping and dependency effects inherited from interval arithmetic. In the following, we suggest the use of DC programming as an alternative. The effectiveness of the proposed scheme is demonstrated in simulation for the considered Li-ion model.
High shares of variable renewable energy necessitate substantial energy storage capacity. However, it remains unclear how to design a market that, on the one hand, ensures a stable and sufficient income for storage fi...
详细信息
Neural generative modelling of sketches has been an active research direction. SketchRNN set a milestone with their sequence-to-sequence variational autoencoder architecture being able to generate hand drawn sketches ...
Neural generative modelling of sketches has been an active research direction. SketchRNN set a milestone with their sequence-to-sequence variational autoencoder architecture being able to generate hand drawn sketches in various classes by modelling them as sequences of displacements between consecutive stroke points. The diversity and variety in the set of handwritten Chinese characters makes them a good candidate for such generative modelling enabling their unconditional generation. However, modelling them as sequences of points causes coarse looking strokes and much longer sequences, consequently requiring use of polygonal approximation algorithms to cut down on points. Instead, we propose and investigate the modelling of Chinese characters as sequences of Bézier curves using the SketchRNN architecture with a few modifications, to allow the model to directly generate smooth curves. This way the encoded representation is smaller while more of the stroke’s characteristics are retained and the generated characters are truly scalable. We also suggest the appropriate preprocessing strategy for the KanjiVG dataset to make it suitable for this purpose. Qualitative evaluation of the results suggests the model demonstrates generation of characters with mostly well-structured and ordered strokes. This was substantiated by quantitative evaluation based on the FrFréchetchet Inception Distance score.
暂无评论