Traumatic brain injury (TBI) assessment is crucial for protecting the health of casualties involved in sudden head-impact incidents. However, traditional methods for assessing TBI are primarily based on bulky imaging equipment, which suffers from time lag, leading to misjudgment of the injury and missing the golden opportunity for treatment. Moreover, most existing Internet of Things (IoT)-aided sensor-based head impact detection studies monitor only a single biomechanical parameter, either the external head impact force or the head center-of-mass acceleration, thus offering an incomplete injury profile. This study proposed a method for rapidly acquiring on-site time-series data to support TBI assessment, which employed flexible piezoelectric sensors to detect head impact, whose electromechanical coupling was characterized by an equivalent-impedance model identified with the genetic algorithm (GA). The voltage signal from the sensor was transmitted to mobile devices via Bluetooth to estimate head-impact biomechanics. Specifically, the impact force on the wearer’s head was converted from the voltage using the sequential quadratic programming algorithm (SQP). Based on the impact force data, the acceleration of the center of mass of the head was predicted with a neural network containing the long short-term memory (LSTM) layer. The experiments were carried out on a head-neck dummy subjected to different levels of impact. The resulting models achieved coefficients of determination (R2) of 0.935 for impact force (amplitude of 0–3 kN) and 0.848 for acceleration of the center of mass (amplitude of 0–120 g). The fabricated wireless flexible wearable sensor was installed on the dummy for performance validation, successfully demonstrating collision detection, impact force and acceleration estimation for TBI assessment, which can be conveniently deployed in helmets used in construction, sports, and emergency rescue, leveraging IoT technology for real-time dual-parameter head-impact estimation and medical-data synchronization to safeguard the golden treatment window and hopefully promote the efficient allocation of medical resources.

Spiking Neural Networks (SNNs) are biologically inspired models that process visual inputs over multiple time steps. However, they often struggle with limited feature discrimination along the temporal dimension due to inherent spatiotemporal invariance. This limitation arises from the redundant activation of certain regions and shared supervision for multiple time steps, constraining the network’s ability to adapt and learn diverse features. To address this challenge, we propose a novel Temporal-Self-Erasing (TSE) supervision method that dynamically adapts the learning regions of interest for different time steps. The TSE method operates by identifying highly activated regions from predictions across multiple time steps and adaptively suppressing them during model training, thereby encouraging the network to focus on less activated yet potentially informative regions. This approach not only enhances the feature discrimination capability of SNNs but also facilitates more effective multi-time-step inference by exploiting more semantic information. Experimental results on benchmark datasets demonstrate that our TSE method significantly improves the classification accuracy and robustness of SNNs.

Spiking Neural Networks (SNNs), with their biologically inspired spatio-temporal dynamics and spike-driven processing, are emerging as a promising low-power alternative to traditional Artificial Neural Networks (ANNs). However, the complex neuronal dynamics and non-differentiable spike communication mechanisms in SNNs present substantial challenges for efficient training. By analyzing the membrane potentials in spiking neurons, we found that their distributions can increasingly deviate from the firing threshold as time progresses, which tends to cause diminished backpropagation gradients and unbalanced optimization. To address these challenges, we propose Deep Temporal-Aligned Gradient Enhancement (DeepTAGE), a novel approach that improves optimization gradients in SNNs from both internal surrogate gradient functions and external supervision methods. Our DeepTAGE dynamically adjusts surrogate gradients in accordance with the membrane potential distribution across different time steps, enhancing their respective gradients in a temporal-aligned manner that promotes balanced training. Moreover, to mitigate issues of gradient vanishing or deviating during backpropagation, DeepTAGE incorporates deep supervision at both spatial (network stages) and temporal (time steps) levels to ensure more effective and robust network optimization. Importantly, our method can be seamlessly integrated into existing SNN architectures without imposing additional inference costs or requiring extra control modules. We validate the efficacy of DeepTAGE through extensive experiments on static benchmarks (CIFAR10, CIFAR100, and ImageNet-1k) and a neuromorphic dataset (DVS-CIFAR10), demonstrating significant performance improvements.

With the rise of "Metaverse" and "Web 3.0", Non-Fungible Token (NFT) has emerged as a kind of pivotal digital asset, garnering significant attention. By the end of March 2024, more than 1.7 billion NFTs have been minted across various blockchain platforms. To effectively locate a desired NFT, conducting searches within a vast array of NFTs is essential. The challenge in NFT retrieval is heightened due to the high degree of similarity among different NFTs, regarding regional and semantic aspects. In this paper, we will introduce a benchmark dataset named "NFT Top1000 Visual-Text Dataset" (NFT1000), containing 7.56 million image-text pairs, and being collected from 1000 most famous PFP1 NFT collections2 by sales volume on the Ethereum blockchain. Based on this dataset and leveraging the CLIP series of pre-trained models as our foundation, we propose the dynamic masking fine-tuning scheme. This innovative approach results in a 7.4\% improvement in the top1 accuracy rate, while utilizing merely 13\% of the total training data (0.79 million vs. 6.1 million). We also propose a robust metric Comprehensive Variance Index (CVI) to assess the similarity and retrieval difficulty of visual-text pairs data. The dataset will be released as an open-source resource. For more details, please refer to: https://github.com/ShuxunoO/NFT-Net.git.

Reverberation is one of the most heavily used effects in music. A new watermarking method based on reverberation theory is proposed in this paper. The embed process same as the process that adds reverberation to recordings in music studio to make it sound quite natural like in a real concert hall. The watermark are embedded by convoluting the original audio signal and a room impulse response that obtained by image method. The parameters for calculating the room impulse response are used as secret key that must be used in the decoding process to ensure security. When decoding, homomorphic deconvolution method is used to separate the kernel function from the cepstrum of the watermarked audio signal, and then take cross-correlation with the delay sequence of room impulse response to get watermark by searching the peak.

Based on the two-dimensional harmonic model, this paper studies the problem of estimating the frequencies of closely spaced complex exponentials in the presence of colored noise, and presents a new estimation approach using the two-dimensional decimation technique. By using the capability of decimation in the time domain to increase the frequency intervals, the proposed method separates the frequencies in the frequency domain, and gives the exact frequency estimation by using the cumulant based matrix pencil method. The proposed method is easy to realize and has successfully improved the performance of the existing two-dimensional frequency estimation methods in the case of closely spaced frequencies. Simulations are provided to show its performance.

The complex signal of two dimensional harmonics is common. Pairing steps are always needed when we estimate the frequency pairs of harmonics which are described by complex signals. We introduce the hypercomplex signal, and use it to study two dimensional harmonics. First, we construct the hypercomplex signal using the signal's original two dimensional harmonics and its Hilbert transform. Then, we present our algorithm for estimating the frequencies of the hypercomplex signal through taking advantage of the properties of Hamilton's quaternion. Some simulations illustrate the prospect of using hypercomplex signals in estimating the parameters of two dimensional harmonics without pairing steps.

The problem of concern here is parameter estimation of harmonics in the presence of cubic nonlinear coupling. Cyclic statistics are employed to estimate the frequencies of harmonics in mutually correlative multiplicative noise of any mean, which is independent of additive noise of any mean. We define a special slice of the sixth-order time-average moment spectrum. It can be applied to obtain the coupled and coupling frequencies. This method needn't constrain the distribution and color of noise. Simulation examples illustrate the algorithms.

For the problem of two-dimensional (2-D) cubic nonlinear phase coupling estimation, special fourth-order cumulant and corresponding slice spectrum of fourth-order cumulant are defined. The slice spectrum of fourth-order cumulant approach is proposed for estimating two-dimensional cubic nonlinear phase coupling. As the fourth-order cumulant of a complex process has 2 different definitions, two definitions are used to extract the coupled and coupling frequency components in Gaussian noise , respectively. Hence the nonlinearity of system is detected effectively. Simulation examples are presented to prove the results of this paper.

A special fourth-order time-average moment spectrum approach is proposed for estimating two-dimensional (2-D) cubic nonlinear coupling. It can be applied to obtain the coupled and coupling frequencies in the noise that multiplicative and additive noises of any mean are mutually independent, and the multiplicative noises are correlative. In the meantime, this method needn't constrain the distribution and the color of noises. Simulation examples illustrate the algorithms.

In this paper, we consider problem about two-dimension (2-D) direction estimation of cyclostationary sources when the received signal is corrupted by both multiplicative noise and additive noise. A 2-D direction finding algorithm based on third-order cyclic moment is proposed to detect nonGaussian cyclostationary sources in the presence of complex noises. Under the cases of existing interference and different distributed noises, the simulation results indicate that the proposed method can remove cyclostationary interference with different cyclic frequency from the desired signals. Furthermore, it can effectively suppress additive and nonzero multiplicative stationary noises with any distribution, no matter they are Gaussian or nonGaussian, white or colored.

Cyclic higher order statistics (CHOS) have powerful signal detection capability and show potential in most mobile communication systems. In this paper, we propose a CMOS-based algorithm to estimate azimuth/elevation using a selected high-resolution planar array to detect nonGaussian coherent cyclostationary sources by designing a fourth-order cyclic cumulant matrix and using eigen vector (EV) approach. For coherent sources caused by multipath propagation, a special forward/backward averaging method is utilized in two-dimension (2-D) case to resolve directions. Simulation results show that, in multipath situation, our algorithm has advantages of suppressing additive stationary noise and Gaussian noise with unknown spatial characteristics, as well as its signal selectivity to remove interferences with different cycle frequencies from the signals of interests (SOI's). Also, array aperture is extended by exploiting a special array geometry.

This paper presents un approach for beamforming based on third-order cyclic statistics. The approach has fairly well accuracy. Further, a new scheme joint beamforming and space-time block coding is suggested that multiple antenna arrays transmit at base station and multi-antenna elements receive at mobile terminal. This scheme improves the receiving error performance of mobile terminal and increases the capacity of the entire wireless communication system. The computer simulation validates that the aforementioned scheme is superior and effective.

This paper addresses the harmonic retrieval problem in colored linear non-Gaussian noise of unknown covariance and unknown distribution. The assumptions made in the reported studies, that the non-Gaussian noise is asymmetrically distributed and no quadratic phase coupling occurs, are released. Using the elaborately defined fourth-order cumulants of the complex noisy observations which are obtained through the Hilbert transform, we can either estimate the noise correlation nonparametrically via cumulant projections or obtain the AR polynomial of the non-Gaussian noise parametrically through ARMA modeling, then it is shown that the prewhitening or prefiltering techniques can be employed to retrieve harmonics respectively. Simulation results are presented to demonstrate the performance of the proposed algorithms.