Mixed-radix Fast Fourier Transform (FFT) algorithms play a crucial role in advanced communication systems such as 5G. However, in the conventional decimation-in-time (DIT) FFT algorithm, the input time-domain sequence...
详细信息
ISBN:
(数字)9798331502768
ISBN:
(纸本)9798331502775
Mixed-radix Fast Fourier Transform (FFT) algorithms play a crucial role in advanced communication systems such as 5G. However, in the conventional decimation-in-time (DIT) FFT algorithm, the input time-domain sequence and intermediate-stage data are not arranged in a natural order, creating a mismatch between the processing element (PE) computation speed and memory bandwidth. Additionally, multiple redundant interstage twiddle factor (TF) generation units operate in parallel to supply TFs, leading to inefficiencies in both computational and storage resources. This article introduces a flexible and reconfigurable architecture that utilizes a mixed-radix FFT approach, supporting FFT sizes from 8 to 4096 points. To enhance hardware efficiency, a CGR-based multi-parallel changeable-radix butterfly unit (BU) is incorporated. Moreover, a conflict-free memory access mechanism, featuring an optimized address generation scheme, is designed around the PE array. A TF sharing and compression technique is also employed to reduce both the area and latency of the TF generation units.
In the field of biomedical engineering, surface electromyography (sEMG) is a key tool for monitoring muscle activity and is widely used in various fields such as human- computer interfaces, muscle fatigue assessment, ...
详细信息
ISBN:
(数字)9798331522742
ISBN:
(纸本)9798331522759
In the field of biomedical engineering, surface electromyography (sEMG) is a key tool for monitoring muscle activity and is widely used in various fields such as human- computer interfaces, muscle fatigue assessment, and rehabilitation training. However, sEMG signals are often affected by power line interference and other noise sources during the acquisition process, which may mask useful information. In this study, a new method combining the variational mode decomposition (VMD) and the crown porcupine optimization (CPO) algorithm, named CPO-VMD, is proposed, aiming to optimize the VMD parameters to improve the denoising effect of sEMG signals. By automatically adjusting the key parameters of the VMD through the intelligent algorithm, this method solves the problem of difficult parameter selection, furthermore, enhancing the denoising efficiency. In this paper, a simulated sEMG signal is constructed and the VMD parameters are optimized by CPO. The experimental results show that the method effectively improves the quality of sEMG signals and provides a more accurate idea for signalprocessing in muscle fatigue assessment and rehabilitation applications.
Emotions express human’s attitude towards external things and play an important role in social life. Automatic emotion recognition technology plays an important role in the field of human-computer interaction. Facial...
详细信息
ISBN:
(数字)9798350375107
ISBN:
(纸本)9798350375114
Emotions express human’s attitude towards external things and play an important role in social life. Automatic emotion recognition technology plays an important role in the field of human-computer interaction. Facial expression analysis is the most intuitive, convenient and effective method for human to perceive emotions. However, fuzzy facial information in large scenes may lead to decreased recognition accuracy, and gait information containing emotional features is less limited by distance to a certain extent
[1]
. In view of the above situation, this paper integrates the gait and facial information and conducts in-depth research on the emotion recognition method based on advanced deep learning algorithm to furtherly improve the recognition accuracy. Aiming at the issue of the relatively limited sample quantity in the gait dataset, from the perspective of spatial graph convolution in the field of graph signalprocessing, a Shift spatio-temporal graph convolution gait emotion recognition network based on the self-attention mechanism is put forward. This effectively mitigates the drawback that the lightweight Shift spatio-temporal graph convolution network disregards the inherent connections among joints and significantly enhances the accuracy of gait emotion recognition.
Rolling element bearing generates a complex vibration due to geometrical flaws, surface irregularities during production, faulty bearing used, and error in the associated component. These vibration signals are usually...
详细信息
A smart, low-cost, stand-alone, and reliable touchless human-computer interaction (HCI) using a tiny 60 GHz radar sensor and a small Raspberry Pi microcomputer with advanced radar signalprocessing and machine learnin...
详细信息
ISBN:
(纸本)9781665492331
A smart, low-cost, stand-alone, and reliable touchless human-computer interaction (HCI) using a tiny 60 GHz radar sensor and a small Raspberry Pi microcomputer with advanced radar signalprocessing and machine learning (ML), is proposed and implemented. Three feature extraction techniques and various ML algorithms are explored and compared. Smart gesture classification is realized by an ensemble ML model with 3 non-deep and deep learning (DL) algorithms and the majority voting scheme to improve classification accuracy. Both radar signalprocessing and ML algorithms are implemented in Python language in a Raspberry Pi microcomputer with an optimization by TensorFlow Lite, which reduces DL CPU time by 120 times. The proposed HCI is also integrated with a menu driven use case, which could be easily adapted to many other gesture control applications for home and industrial appliances.
Sleep apnea is a sleep-related condition characterized by the swelling or relaxation of throat muscles, causing a blockage in the upper airways. This obstruction interrupts normal breathing patterns during sleep. The ...
详细信息
ISBN:
(数字)9798350384369
ISBN:
(纸本)9798350384376
Sleep apnea is a sleep-related condition characterized by the swelling or relaxation of throat muscles, causing a blockage in the upper airways. This obstruction interrupts normal breathing patterns during sleep. The current gold standard polysomnography (PSG) test for detecting sleep apnea is expensive, inconvenient, and lacks widespread availability for the general population. This underscores the need for more user-friendly and readily available solutions to diagnose sleep apnea. Collaborative efforts between Deep learning practitioners and healthcare professionals are essential to ensure that the developed techniques align with established clinical standards and contribute effectively to early diagnosis, prediction, and management of sleep apnea. This paper delves into recent studies on intelligent sleep apnea detection mechanisms. We emphasize the background of sleep apnea and the evolution of medical understanding and technological advancement. We demonstrated the diagnosis of sleep apnea through experts like medical professionals. We delve into recent studies on intelligent sleep apnea detection mechanisms using numerous deep-learning techniques and methodologies. In the transition, our attention is directed toward recent research that underscores the significance of integrating health-related features through classifier training.
Object detection is a fundamental task in computer vision, consisting of both classification and localization tasks. Previous works mostly perform classification and localization with shared feature extractor like Con...
详细信息
Object detection is a fundamental task in computer vision, consisting of both classification and localization tasks. Previous works mostly perform classification and localization with shared feature extractor like Convolution Neural Network. However, the tasks of classification and localization exhibit different sensitivities with regard to the same feature, hence the "task spatial misalignment" issue. This issue can result in a hedge issue between the performances of localizer and classifier. To address these issues, we first propose a novel Dynamic Coefficient Loss to simultaneously consider and balance the performances of classification and localization tasks. To well address anchor label misjudgment issue in irregular- shaped object detection, we define a new classification-aware IoU metric to assign anchors intelligently. Finally, we further introduce the localization factor into NMS by proposing a Classification-Localization balanced NMS. Extensive experiments on MS COCO and PASCAL VOC demonstrate that our proposals can improve RetinaNet by around 1.5% AP with various backbones.
R wave, as an important reference for determining each band of ECG signal, is the premise of ECG automatic analysis. In view of the problem that the preprocessing process of most R-wave recognition algorithms affects ...
详细信息
ISBN:
(纸本)9781665468138
R wave, as an important reference for determining each band of ECG signal, is the premise of ECG automatic analysis. In view of the problem that the preprocessing process of most R-wave recognition algorithms affects the recognition accuracy, a new R-wave recognition method based on the combination of wavelet transform and adaptive threshold method is proposed, which improves the stability and accuracy of R-wave recognition. Firstly, the ECG signal is decomposed based on wavelet, and the 2 3 rd layer wavelet signal is obtained. Then the sliding window width and initial threshold are set. The threshold is adjusted according to the signal situation and the window width is increased to quickly identify the $R$ wave. Finally, the simulation results show that the algorithm is effective in R-wave recognition of heart rate variability ECG signal during shooting.
Biomedical signals provide valuable data on the activities of many bodily systems. Biomedical signals are often non-stationary and exhibit non-linear behaviour. Hence, it is rather challenging to get useful data from ...
详细信息
ISBN:
(数字)9798331518981
ISBN:
(纸本)9798331518998
Biomedical signals provide valuable data on the activities of many bodily systems. Biomedical signals are often non-stationary and exhibit non-linear behaviour. Hence, it is rather challenging to get useful data from these signals alone by observation in the temporal domain. Because of this, signalprocessing methods are used to extract crucial aspects from these signals in order to diagnose various illnesses. Automated methods that do not need human intervention and have a high rate of anomaly detection may help doctors with the diagnosis and treatment of a variety of ailments. Back in the day, machine learning was the go-to approach for bio signal analysis. Deep learning algorithms are currently being developed. Electrocardiogram (ECG) signals indicate that the ANS, which regulates the heart's regular rhythm, is in a state of active regulation. This ECG signal is non-stationary and non-linear, much like any other bio signal. An electrocardiogram (ECG) may reveal symptoms of a number of illnesses. There were significant variations in the examined models' accuracy, recall, precision, and F1 score, and CNN 1-RNN, CNN 1-LSTM, and CNN 1-GRU all performed differently. While CNN 2-RNN and CNN2-LSTM models showed clear advantages and disadvantages, CNN2-GRU showed a more even distribution of the two. The CNN7 Layer model nevertheless managed to show impressive results, even with these modifications. On the other hand, CNN 7-LSTM comes out as the most effective approach, with impressive results: 93.5% accuracy.
Traditional methods of monitoring vital signs in vehicles often fall short when it comes to response time, especially in fast-moving situations. To tackle this, our paper introduces a groundbreaking solution: leveragi...
详细信息
ISBN:
(数字)9798331508456
ISBN:
(纸本)9798331508463
Traditional methods of monitoring vital signs in vehicles often fall short when it comes to response time, especially in fast-moving situations. To tackle this, our paper introduces a groundbreaking solution: leveraging millimeter-wave radar technology to monitor vital signs in real time within the vehicle, providing timely alerts when necessary. This approach not only enhances the accuracy of readings but also ensures a swift response that is crucial during vehicle operation. By tapping into the unique abilities of millimeter-wave radar to detect subtle human signs, we propose a novel method that processes radar signals to monitor vital signs like heart rate and respiratory rate. This technique is highly effective in meeting the stringent demands of real-time accuracy and speed, especially while driving. Our system includes three main components: radar sensors, a signalprocessing unit, and an early-warning control module, working together seamlessly to ensure optimal monitoring. The radar sensor continuously transmits and receives millimeter-wave signals, allowing it to track vital signs without interruption. A specially designed feature extraction algorithm filters out noise and extracts critical information, including heart and respiratory rates. Then, advancedalgorithms analyze this data in real time, spotting any abnormalities and triggering alerts through the control module. What sets this system apart is its rigorous testing under various vehicle speeds and environmental conditions, confirming its stability, reliability, and fast response times. In real-world applications, the system responds to a high heart rate in 3.7 to 4.0 seconds and to a rapid breathing rate in 4.1 to 4.3 seconds. This millimeter-wave radar-based system offers a fresh approach to vehicle safety, combining practicality with promising future potential.
暂无评论