Search

The high-accuracy detection of multiple distinct human emotions is crucial for advancing affective computing, mental health diagnostics, and human–computer interaction. The integration of deep learning networks with entropy measures holds significant potential in neuroscience and medicine, especially for analyzing EEG-based emotion states. This study proposes a method combining ResNet18 with differential entropy to identify five types of human emotions (happiness, sadness, fear, disgust, and neutral) from EEG signals. Our approach first calculates the differential entropy of EEG signals to capture the complexity and variability of the emotional states. Then, the ResNet18 network is employed to learn feature representations from the differential entropy measures, which effectively captures the intricate spatiotemporal dynamics inherent in emotional EEG patterns using residual connections. To validate the efficacy of our method, we conducted experiments on the SEED-V dataset, achieving an average accuracy of 95.61%. Our findings demonstrate that the combination of ResNet18 with differential entropy is highly effective in classifying multiple distinct human emotions from EEG signals. This method shows robust generalization and broad applicability, indicating its potential for extension to various pattern recognition tasks across different domains. [ABSTRACT FROM AUTHOR]

The construction of complex networks from electroencephalography (EEG) proves to be an effective method for representing emotion patterns in affection computing as it offers rich spatiotemporal EEG features associated with brain emotions. In this paper, we propose a novel method for constructing complex networks from EEG signals for emotion recognition, which begins with phase space reconstruction to obtain ordinal patterns and subsequently forms a graph network representation from the sequence of ordinal patterns based on the visibility graph method, named ComNet-PSR-VG. For the proposed ComNet-PSR-VG, the initial step involves mapping EEG signals into a series of ordinal partitions using phase space reconstruction, generating a sequence of ordinal patterns. These ordinal patterns are then quantified to form a symbolized new sequence. Subsequently, the resulting symbolized sequence of ordinal patterns is transformed into a graph network using the visibility graph method. Two types of network node measures, average node degree (AND) and node degree entropy (NDE), are extracted from the graph networks as the inputs of machine learning for EEG emotion recognition. To evaluate the effectiveness of the proposed construction method of complex networks based on the visibility graph of ordinal patterns, comparative experiments are conducted using two types of simulated signals (random and Lorenz signals). Subsequently, EEG emotion recognition is performed on the SEED EEG emotion dataset. The experimental results show that, with AND as the feature, our proposed method is 4.88% higher than the existing visibility graph method and 12.23% higher than the phase space reconstruction method. These findings indicate that our proposed novel method for constructing complex networks from EEG signals not only achieves effective emotional EEG pattern recognition but also exhibits the potential for extension to other EEG pattern learning tasks, suggesting broad adaptability and application potential for our method. [ABSTRACT FROM AUTHOR]

With the rapid development of technologies like artificial intelligence, high-performance computing chips are playing an increasingly vital role. The inverse hyperbolic sine and inverse hyperbolic cosine functions are of utmost importance in fields such as image blur and robot joint control. Therefore, there is an urgent need for research into high-precision, high-performance hardware Intellectual Property (IP) for arcsinh and arccosh functions. To address this issue, this paper introduces a novel 128-bit low-latency floating-point hardware IP for arcsinh and arccosh functions, employing an enhanced Coordinate Rotation Digital Computer (CORDIC) algorithm, achieving a computation precision of 113 bits in just 32 computation cycles. This significantly enhances computational efficiency while reducing hardware implementation latency. The results indicate that, when compared to Python standard results, the calculated error of the proposed hardware IP does not exceed 8 × 10 − 34 . Furthermore, this paper synthesizes the completed IP using the TSMC 65 nm process, with a total IP area of 2.1056 mm 2 . Operating at a frequency of 300 MHz, its power is 22.4 mW. Finally, hardware implementation and resource analysis are conducted and compared on an Field Programmable Gate Array (FPGA). The results show that the improved algorithm trades a slight area increase for lower latency and higher accuracy. The designed hardware IP is expected to provide a more accurate and efficient computational tool for applications like image processing, thereby advancing technological development. [ABSTRACT FROM AUTHOR]

According to the main construction procedures of underground caverns, an analysis model for the excavation of a certain pumping station underground cavern group was established, and its excavation and support process are dynamically simulated. The calculation schemes of unsupport and support are compared with indicators such as plastic zone, bolt axial force and displacement field distribution. The analysis results show that after the support, the plastic zone distribution and deformation characteristics of the rock around the cave are greatly improved. The bolt axial force also meets the requirements, indicating that the support parameters and support methods are reasonable. The supporting effect has been well simulated and realized.

In the field of digital signal processing, such as in navigation and radar, a significant number of high-precision arctangent function calculations are required. Lookup tables, polynomial approximation, and single/double-precision floating-point Coordinate Rotation Digital Computer (CORDIC) algorithms are insufficient to meet the demands of practical applications, where both high precision and low latency are essential. In this paper, based on the concept of trading area for speed, a four-step parallel branch iteration CORDIC algorithm is proposed. Using this improved algorithm, a 128-bit quad-precision floating-point arctangent function is designed, and the hardware circuit implementation of the arctangent algorithm is realized. The results demonstrate that the improved algorithm can achieve 128-bit floating-point arctangent calculations in just 32 cycles, with a maximum error not exceeding 2 × 10 − 34 rad. It possesses exceptionally high computational accuracy and efficiency. Furthermore, the hardware area of the arithmetic unit is approximately 0.6317 mm 2 , and the power consumption is about 40.6483 mW under the TSMC 65 nm process at a working frequency of 500 MHz. This design can be well suited for dedicated CORDIC processor chip applications. The research presented in this paper holds significant value for high-precision and rapid arctangent function calculations in radar, navigation, meteorology, and other fields. [ABSTRACT FROM AUTHOR]

The presence of a large amount of echoes significantly impairs the quality and intelligibility of speech during communication. To address this issue, numerous studies and models have been conducted to cancel echo. In this study, we propose a multi-stage acoustic echo cancellation model that utilizes an adaptive filter and a deep neural network. Our model consists of two parts: the Speex algorithm for canceling linear echo, and the multi-scale time-frequency UNet (MSTFUNet) for further echo cancellation. The Speex algorithm takes the far-end reference speech and the near-end microphone signal as inputs, and outputs the signal after linear echo cancellation. MSTFUNet takes the spectra of the far-end reference speech, the near-end microphone signal, and the output of Speex as inputs, and generates the estimated near-end speech spectrum as output. To enhance the performance of the Speex algorithm, we conduct delay estimation and compensation to the far-end reference speech. For MSTFUNet, we employ multi-scale time-frequency processing to extract information from the input spectrum. Additionally, we incorporate an improved time-frequency self-attention to capture time-frequency information. Furthermore, we introduce channel time-frequency attention to alleviate information loss during downsampling and upsampling. In our experiments, we evaluate the performance of our proposed model on both our test set and the blind test set of the Acoustic Echo Cancellation challenge. Our proposed model exhibits superior performance in terms of acoustic echo cancellation and noise reverberation suppression compared to other models. [ABSTRACT FROM AUTHOR]

In the sensorless control algorithm of permanent magnet synchronous motors (PMSMs), there is a significant research focus on developing a high-performance position observer and a cost-effective, high-performance processor specifically designed for PMSMs. In this paper, we first reviewed the current mainstream position observation algorithms. Subsequently, we proposed a position observation algorithm based on a fixed-point nonlinear flux linkage model that introduces a variable gain model and an observer compensation model. Furthermore, we implemented the improved position observation algorithm in the circuit and designed a System-on-Chip (SOC) suitable for the sensorless control of high-speed permanent magnet synchronous motors using the ARM Cortex-M0 core. The functionality of the proposed SOC was verified on the Field Programmable Gate Array (FPGA) platform. The results demonstrate that the improved algorithm offers an excellent angle observation accuracy and fast convergence speed. The error between the observed angle and the actual angle is less than 0.2 rad. Additionally, the designed SOC balances the high efficiency of hardware implementation with the flexibility of software implementation. When running a sensorless control algorithm for PMSM, it achieves a 30.3% increase in execution efficiency, enabling support for higher frequency output PWM switching and making it suitable for the high-precision control of high-speed PMSM. [ABSTRACT FROM AUTHOR]

Artificial intelligence has revolutionized image and speech recognition, but the neural network fitting method has limitations. Neuromorphic chips that mimic biological neurons can better simulate the brain's information processing mechanism. As the basic computing component of the new neuromorphic network, the new neural computing unit's design and implementation have important significance; however, complex dynamical features come with a high computational cost: approximate computing has unique advantages, in terms of optimizing the computational cost of neural networks, which can solve this problem. This paper proposes a hardware implementation of an approximate spiking neuron structure, based on a simplified piecewise linear model (SPWL), to optimize power consumption and area. The proposed structure can achieve five major neuron spiking generation patterns. The proposed design was synthesized and compared to similar designs, to evaluate its potential advantages and limitations. The results showed that the approximate spiking neuron had the lowest computational cost and the fastest computation speed. A typical spiking neural network was constructed, to test the usability of the SPWL model. The results showed that the proposed approximate spiking neuron could work normally in the spiking neural network, and achieved an accuracy of 94% on the MNIST dataset. [ABSTRACT FROM AUTHOR]

During 3D rendering by GPU, jagged patterns will appear at the edge of the image due to the lack of sampling points. To increase the continuity of the image, we use an anti-aliasing technique to process these jagged patterns. In this paper, we review the current anti-aliasing techniques and propose a multi-parallel anti-aliasing algorithm based on Morphological Anti-Aliasing (MLAA). Through an experiment and a comparison of the results, we find that our algorithm achieves better anti-aliasing performance at the edge of the image than Multiple Sampling Anti-Aliasing (MSAA) and MLAA by setting more color gradients. Moreover, our algorithm consumes much less time than MLAA in FPGA implementation by performing edge detection of the image, classification of the aliasing patterns, acquisition of the blend weight coefficients, and calculation of the anti-aliased color in three channels of the image simultaneously. In addition, our algorithm also consumes much less memory than MLAA in FPGA implementation by scaling and optimizing the area texture used by MLAA. A comparison of related works indicates that it is more favorable for the proposed algorithm to perform the anti-aliasing of the image than MSAA and MLAA. [ABSTRACT FROM AUTHOR]

Histopathological image analysis is an examination of tissue under a light microscope for cancerous disease diagnosis. Computer-assisted diagnosis (CAD) systems work well by diagnosing cancer from histopathology images. However, stain variability in histopathology images is inevitable due to the use of different staining processes, operator ability, and scanner specifications. These stain variations present in histopathology images affect the accuracy of the CAD systems. Various stain normalization techniques have been developed to cope with inter-variability issues, allowing standardizing the appearance of images. However, in stain normalization, these methods rely on the single reference image rather than incorporate color distributions of the entire dataset. In this paper, we design a novel machine learning-based model that takes advantage of whole dataset distributions as well as color statistics of a single target image instead of relying only on a single target image. The proposed deep model, called stain acclimation generative adversarial network (SA-GAN), consists of one generator and two discriminators. The generator maps the input images from the source domain to the target domain. Among discriminators, the first discriminator forces the generated images to maintain the color patterns as of target domain. While second discriminator forces the generated images to preserve the structure contents as of source domain. The proposed model is trained using a color attribute metric, extracted from a selected template image. Therefore, the designed model not only learns dataset-specific staining properties but also image-specific textural contents. Evaluated results on four different histopathology datasets show the efficacy of SA-GAN to acclimate stain contents and enhance the quality of normalization by obtaining the highest values of performance metrics. Additionally, the proposed method is also evaluated for multiclass cancer type classification task, showing a 6.9% improvement in accuracy on ICIAR 2018 hidden test data. [ABSTRACT FROM AUTHOR]

Floating-point division is indispensable and becoming increasingly important in many modern applications. To improve speed performance of floating-point division in actual microprocessors, this paper proposes a low-latency architecture with a multi-precision architecture for floating-point division which will meet the IEEE-754 standard. There are three parts in the floating-point division design: pre-configuration, mantissa division, and quotient normalization. In the part of mantissa division, based on the fast division algorithm, a Predict–Correct algorithm is employed which brings about more partial quotient bits per cycle without consuming too much circuit area. Detailed analysis is presented to support the guaranteed accuracy per cycle with no restriction to specific parameters. In the synthesis using TSMC, 90 nm standard cell library, the results show that the proposed architecture has ≈63.6% latency, ≈30.23% total time (latency × period), ≈31.8% total energy (power × latency × period), and ≈44.6% efficient average energy (power × latency × period/efficient length) overhead over the latest floating-point division structure. In terms of latency, the proposed division architecture is much faster than several classic processors. [ABSTRACT FROM AUTHOR]

In this study, a novel human acoustic perception motivated Wiener filter speech enhancement system is presented to cope with real‐world interfering background noises. Guiding by the speech presence probability, two alternative methods are proposed to reduce the noise by adopting the audible sound pressure level (SPL) and the masking characteristic of the human auditory system to achieve better listening comfort level. More specifically, when the probability of speech presence in the noisy signal is less than the decision threshold, a new SPL compressed method effectively reduces the noise. When the speech presence probability is more than the decision threshold, an improved acoustical mask threshold constrained Wiener filter approach enhances the noisy speech. Moreover, in order to evaluate the performance of the new system, the proposed algorithm is compared with the classic prior signal‐to‐noise ratio‐based Wiener filter and three acoustic perception related algorithms. The experimental results show that the proposed algorithm significantly outperforms the four comparing algorithms in terms of speech quality and intelligibility either in stationary or moderate non‐stationary noisy environments. Thus, the intended approach can be employed as the front‐end module for various speech‐related applications. [ABSTRACT FROM AUTHOR]

Electroencephalography (EEG) signal is an electrophysiological recording from electrodes placed on the scalp to reflect the electrical activities of the brain. Auditory brainstem response (ABR) is one type of EEG signals in response to an auditory stimulus, and it has been widely used to evaluate the potential disorders of the auditory function within the brain. Currently, the ABR measurements in the clinic usually adopt a fixed stimulation rate (FSR) technique in which the late evoked response could contaminate the ABR signals and deteriorate the waveform differentiation after averaging, thus compromising the overall auditory function assessment task. To resolve this issue, this study proposed a random stimulation rate (RSR) method by integrating a random interval between two adjacent stimuli. The results showed that the proposed RSR method was consistently repeatable and reliable in multiple trials of repeated measurements, and there was a large amplitude of successive late evoked response that would contaminate the ABR signals for conventional FSR methods. The ABR waveforms of the RSR method showed better wave I–V morphology across different stimulation rates and stimulus levels, and the improved ABR morphology played an important role in early diagnoses of auditory pathway abnormities. The correlation coefficients as functions of averaging time showed that the ABR waveform of the RSR method stabilizes significantly faster, and therefore, it could be used to speed up current ABR measurements with more reliable testing results. The study suggests that the proposed method would potentially aid the adequate reconstruction of ABR signals towards a more effective means of hearing loss screening, brain function diagnoses, and potential brain–computer interface. [ABSTRACT FROM AUTHOR]

The performance of speech enhancement algorithms can be further improved by considering the application scenarios of speech products. In this paper, we propose an attention-based branchy neural network framework by incorporating the prior environmental information for noise reduction. In the whole denoising framework, first, an environment classification network is trained to distinguish the noise type of each noisy speech frame. Guided by this classification network, the denoising network gradually learns respective noise reduction abilities in different branches. Unlike most deep neural network (DNN)-based methods, which learn speech reconstruction capabilities with a common neural structure from all training noises, the proposed branchy model obtains greater performance benefits from the specially trained branches of prior known noise interference types. Experimental results show that the proposed branchy DNN model not only preserved better enhanced speech quality and intelligibility in seen noisy environments, but also obtained good generalization in unseen noisy environments. [ABSTRACT FROM AUTHOR]

Measures of predictability in physiological signals based on entropy metrics have been widely used in the application domain of medical assessment and clinical diagnosis. In this paper, we propose a new entropy-based pattern learning by a combination of singular spectrum analysis (SSA) and entropy measures for assessment of physiological signals. Physiological signals are first represented as a series of SSA components, and then well-established entropy measures are extracted from the resulting SSA components that can help to facilitate the features extraction from physiological signals. The entropy measures of notable SSA components are used to form input features and fed into pattern classifier. To demonstrate its validity, applicability, and versatility, the proposed entropy-based pattern learning is used to perform medical assessments with three kinds of classical physiological signals, that is, electroencephalogram (EEG), electromyogram (EMG), and RR-interval signals. Experiments demonstrate that in all cases, the proposed entropy-based pattern learning can effectively capture specific biosignal patterns of physiological signals and achieve excellent identification performances for the assessments of EEG, EMG, and RR-interval signals. Besides, through the comparison of the identification performances for entropy-based pattern learning based on the physiological signals themselves and the SSA components, it is concluded that the discriminating power of entropy-based pattern learning based on the SSA components is much stronger than that based on the physiological signals themselves. Since it can be easily extended to any other physiological signal analysis, the proposed entropy-based pattern learning may use as an efficient approach to reveal biosignal patterns for medical assessment of physiological signals. [ABSTRACT FROM AUTHOR]

The effect of dilute Co doping on the epitaxial growth of GdBCO films was investigated by self-developed fluorine-free polymer-assisted metal-organic deposition (PA-MOD) method. Under the same sintering heat treatment, dilute Co doping dramatically improved the crystallinity, microstructure, and critical current density (Jc) of GdBCO films. High Jc of 4.0 MA/cm2 at 77 K and self-field was obtained in Co-doped film after only sintering for 30 min, which is eight times of the Jc for the undoped film. Also, Co-doped GdBCO film sintered for as short as 20 min still possesses a high Jc of 1.5 MA/cm2, whereas the Jc is 0 for the similarly sintered pure film. These results suggest that dilute Co doping facilitates the epitaxial growth of the high-performance GdBCO film. [ABSTRACT FROM AUTHOR]

In this paper, we apply the Freeman–Durden decomposition method to the compact polarimetric synthetic aperture radar (SAR) interferometry (C-PolInSAR) data. The complex cross-correlation matrix of the single-baseline C-PolInSAR is decomposed into three $2\times 2$ covariance matrices, which correspond to the surface, the dihedral and the volume scattering, respectively. Under reasonable assumptions, each model of the three scattering mechanisms are derived and formulated in term of the $\pi /4$ C-PolInSAR mode and the circular polarization transmitting and linear polarization reception (CTLR) C-PolInSAR mode. Prior to the target decomposition, detailed steps to retrieve the volume scattering amplitude are given, which are remarkably different for the $\pi /4$ mode and the CTLR mode. In addition, the topographic phase is also retrieved based on the random volume over ground (RVoG) model with the C-PolInSAR data. Finally, a numerical method is applied to solve the system of nonlinear equations in the target decomposition procedure, and the amplitude contributions, as well as the vertical phase center heights are retrieved for each of the three scattering mechanisms with the C-PolInSAR data. The proposed target decomposition method has been verified both with a simulated forest data, and with the L-band BioSAR2008 airborne data acquired by the DLR E-SAR system of a forest site in the north of Vindeln municipality, Sweden, wherein detailed performance results for the proposed method are presented as compared to the more usual quad polarimetric SAR interferometry target decomposition. [ABSTRACT FROM AUTHOR]