Quantitative ultrasound (QUS) holds promise for non-invasive placental tissue characterization and disease detection, yet its clinical application is hindered by the effort required for placental segmentation to ident...
详细信息
ISBN:
(纸本)9798350371918;9798350371901
Quantitative ultrasound (QUS) holds promise for non-invasive placental tissue characterization and disease detection, yet its clinical application is hindered by the effort required for placental segmentation to identify a region-of-interest (ROI) for calculation. Manual segmentation is time-consuming and prone to variability due to the irregular boundaries of the placental tissue. This study aims to develop an automatic placental identification and segmentation model to facilitate QUS integration into clinical practice, alleviate clinician workload, and enable real-time feedback on QUS quality. We employed Mask R-CNN within the Detectron2 framework to automate placental segmentation using a dataset of 149 B-mode ultrasound images annotated by a medical image specialist (M.D.) covering various trimesters. Inference on a validation subset (30 images) yielded an average Dice Similarity Coefficient (DSC) of 0.863. Our model gave quality predictions for a majority of the test images with 57% of segmentations achieving DSC >= 0.85 and 40% with DSC >= 0.90. The model's low inference time of approximately 55 ms per iteration supports near real-time QUS processing. Future work will focus on expanding the dataset, optimizing the loss function, and addressing edge effects on QUS to further enhance segmentation quality.
The proceedings contain 157 papers. The topics discussed include: a reversible data hiding scheme for ECG signals using CNN-PEE;design and development of in-memory-compute SRAM cell using 45nm technology;stand alone o...
ISBN:
(纸本)9798350379525
The proceedings contain 157 papers. The topics discussed include: a reversible data hiding scheme for ECG signals using CNN-PEE;design and development of in-memory-compute SRAM cell using 45nm technology;stand alone or non-standalone 5G tactical edge network architecture for military and use case scenarios;secrecy capacity optimization in RF/FSO systems: impact of mixed Rayleigh and log-normal fading with atmospheric turbulence;real-time eye-tracking mouse control system using OpenCV and facial landmark detection;a time-efficient path navigating landmine detection robot;optimizing GPS positioning: a deeplearning approach to improve accuracy;a novel structure of on-chip multilayered half-turn inductor for RF applications;and low profile wideband 3 element parasitic hexagonal patch antenna.
This paper deals with the problem of video-based face recognition. Nowadays, facial recognition methods have made a big step forward, but video-based recognition with its poor quality, difficult lighting conditions, a...
详细信息
This paper deals with the problem of video-based face recognition. Nowadays, facial recognition methods have made a big step forward, but video-based recognition with its poor quality, difficult lighting conditions, and real-time requirements is still a difficult and unfinished *** paper uses the apparatus of convolutional networks for various stages of processing: for capturing and detecting a face, for constructing a feature vector, and finally for recognition. All algorithms are implemented and studied in the Matlab environment to simplify their further export to embedded applications.
Edge intelligence has recently attracted great interest from industry and academia, and it greatly improves the processing speed at the edge by moving data and artificial intelligence to the edge of the network. Howev...
详细信息
Edge intelligence has recently attracted great interest from industry and academia, and it greatly improves the processing speed at the edge by moving data and artificial intelligence to the edge of the network. However, edge devices have bottlenecks in battery capacity and computing power, making it challenging to perform computing tasks in dynamic and harsh network environments. Especially in disaster scenarios, edge (rescue) devices are more likely to fail due to unreliable wireless communications and scattered rescue requests, which makes it urgent to explore how to provide low-latency, reliable services through edge collaboration. In this article, we investigate the task offloading mechanism in mobile edge computing networks, aiming to ensure fault tolerance and rapid response of computing services in dynamic and harsh scenarios. Specifically, we design a fault-tolerant distributed task offloading scheme, which minimizes task execution time and system energy consumption through the multi-agent proximal policy optimization algorithm. Furthermore, we introduce logarithmic ratio reward functions and action masking to reduce the impact of different task queue lengths while accelerating model convergence. Numerical results show that the proposed algorithm is suitable for service failure scenarios, effectively meeting the reliability requirements of tasks while simultaneously reducing system energy consumption and processing latency.
Coastal wetlands are critical from an ecological, hydrological, and biodiversity point of view [1]. Today, these areas are under threat and have been declining every year since they were first studied [2]. Satellite i...
详细信息
ISBN:
(纸本)9798350360332;9798350360325
Coastal wetlands are critical from an ecological, hydrological, and biodiversity point of view [1]. Today, these areas are under threat and have been declining every year since they were first studied [2]. Satellite images are an invaluable tool for understanding wetlands. Thanks to the revisit capability of satellites, coastal wetlands can be mapped in realtime, allowing us to understand how they are changing. This work aims to map these ecosystems as accurately as possible. In order to achieve this, two types of sensors were used: a Sentinel-2 time series, a SPOT6/7 image, and altimetric data (RGE Alti (c)). Two automated learning algorithms were used: random forest (RF) and convolutional neural networks (CNN). Two methods were also used: a pixel approach and an object approach. The results show that the synergy of Sentinel-2 and SPOT with the contribution of indices and an object method works best, with an overall accuracy of 0.90 compared to 0.78 for a SPOT-6 pixel-by-pixel approach.
Flexible tactile sensors have the ability to provide unparalleled levels of tactile sensation, including information regarding roughness, contact force, and contact location. However, it remains a challenge to achieve...
详细信息
Flexible tactile sensors have the ability to provide unparalleled levels of tactile sensation, including information regarding roughness, contact force, and contact location. However, it remains a challenge to achieve precise contact location sensing that is decoupled from sensor strain and touching forces. This paper proposes a novel data-driven approach for force contact location sensing (FCLS) with the influence of sensor strain and forces based on scatter signals (SS) of the ultrasonic waveguide. First, the envelope of the force contact scatter signal (FCSS) is extracted via the Hilbert transform, which retrieves the global features of SS. The time-frequency spectrogram is obtained via continuous wavelet transform, which extracts the local features of SS. Second, a deep convolutional neural network (CNN) is utilized to extract these features separately and concentrate them together. Third, based on the outputs of the CNN, a multilayer perception regression model is applied to acquire the force contact location. The experimental results indicate that the accuracy of the proposed FCLS method has a mean absolute error of 0.627 mm and a mean relative error of 3.19%. This research provides a foundation for further multimodal sensing using ultrasonic waveguides and its application in robotic sensing. This article proposes a novel data-driven approach for force contact location sensing (FCLS), based on scatter signals of the ultrasonic waveguide, with the influence of sensor strain and forces. The experimental results indicate that the accuracy of the proposed FCLS method has an mean absolute error loss of 0.627 mm and a mean relative error loss of 3.19%.image (c) 2024 WILEY-VCH GmbH
In large organizations, the number of financial transactions can grow rapidly, driving the need for fast and accurate multi-criteria invoice validation. Manual processing remains error-prone and time-consuming, while ...
详细信息
This research focuses on the development and evaluation of a robotic grasping system that combines a force sensor and a deep Q-learning algorithm, with objects identified using Mask R-CNN. The system's performance...
详细信息
Lithography defect detection plays a vital role in chip production. A sensitive and reliable defect detection method is very important to ensure the quality of chips. deeplearning has shown promising results in learn...
详细信息
In the rapidly evolving digital age, social media platforms have become central to human interaction, offering a unique and dynamic window into the diverse spectrum of human emotions. Among these platforms, Twitter st...
详细信息
ISBN:
(纸本)9798331540661;9798331540678
In the rapidly evolving digital age, social media platforms have become central to human interaction, offering a unique and dynamic window into the diverse spectrum of human emotions. Among these platforms, Twitter stands out due to its concise and real-time communication format, capturing a wide array of sentiments and emotional expressions. This research utilizes a Twitter dataset to explore the complex human emotions as expressed through tweets. Each entry in the dataset consists of a snippet of a Twitter message categorized into one of six distinctive emotional hues: Sadness (0), Joy (1), Love (2), Anger (3), Fear (4), and Surprise (5). The primary aim of this study is to analyze the emotions of Twitter messages by utilizing advanced techniques for sentiment analysis and emotion classification, specifically focusing on developing a Convolutional Neural Network (CNN) model. The proposed approach involves pre-processing the dataset, training the CNN model, and evaluating its performance using metrics like accuracy, precision, recall, and F1-score. The findings of this study not only enhance our understanding of how emotions are communicated in the digital space but also offer practical insights for applications in marketing, customer service, and mental health.
暂无评论