Single-mode Systems Research Articles

Research on Modelling Capability of English Multimodal File Search based on Transformer

With the exponential growth of file data in the multimedia era, file retrieval ability to achieve effective data management has become a hot research field. Based on people’s English file search needs, this paper proposes an English multimodal file search model based on transformer. Through ablation experiments on two public data sets and comparison experiments with the benchmark model, the effectiveness and superiority of the proposed transformers algorithm model in multi-modal data processing are verified. The multi-modal fusion retrieval system can usually achieve better performance than the single-modal retrieval system. This experiment focuses on three modes: Audio, Image and Text. The experimental results show that the proposed method can not only improve the efficiency of file search, but also extract modal features and perform feature fusion better. In the future, we can further explore different types of other attention mechanisms or integrate a variety of different architectures to further enhance the feasibility and superiority of multimodal file search

Screening of vitrification solution for single-mode electromagnetic rewarming system for organ cryopresesrvation

Screening of vitrification solution for single-mode electromagnetic rewarming system for organ cryopresesrvation

Feature-level fusion of face and speech based multimodal biometric attendance system with liveness detection

The rapid growth of deep learning and the Internet of Things has spurred the need for touchless biometric systems in areas where cleanliness and non-intrusive user interaction are critical. In order to authenticate seamlessly, traditional biometric methods such as fingerprint, hand, etc. recognition require physical contact and, therefore, risk hygiene issues, hence making face and speaker verification more viable alternatives. A robust Multimodal Biometric Attendance System (MBAS) is needed due to the vulnerabilities and limitations of single modality systems. In this research, we introduce MBAS using feature-level fusion of speech data with face data, combining the best of both worlds. The textural features based on a person’s facial appearance are integrated with dynamic speech information for liveness detection, followed by dimensionality reduction using linear discriminant analysis, and then incorporated into a Bi-LSTM classifier. Therefore, for better security, accuracy, and anti-spoofing attacks, this approach is proposed in addition to increasing accuracy as well as enhancing security against spoofing attacks. Two publicly available datasets, DeepfakeTIMIT and AVSpeech, are extensively explored to evaluate different fusion strategies, classifier types, and standard performance metrics. The proposed system outperformed other cutting-edge biometric based systems by exhibiting a 97.51% high accuracy rate with a precision of 99.10% and an equal error rate of 2.48%. These findings affirm the effectiveness and possible real-world applications of the MBAS concept, along with its enhancement ensuring safety. Furthermore, this study underscores the importance of incorporating advanced liveness detection into secure contactless biometrics solutions for modern attendance management in various industries that encompass both face and voice modalities.

Passenger origin–destination estimation in public transit using boarding count data

Passenger boarding and alighting data are critical for understanding travel patterns in transit systems, yet many systems only record boarding, making it difficult to track alighting. This study introduces a method to estimate alighting counts from boarding data using transit symmetry at a cluster level. Bus stops are grouped into clusters, assuming equal inflow and outflow, and a gravity model is calibrated to estimate an origin–destination (OD) matrix. Using data from three bus routes in Regina, Canada (2017), the method achieved a symmetry ratio of 60% across clusters and a mean structural similarity measure score of 0.84, validating the accuracy of the OD estimates. This approach provides a practical solution to the challenge of missing alighting data. While effective for single-mode transit systems, future research will expand its application to multi-modal systems and address factors like seasonal and land-use variations, offering a cost-effective alternative for OD estimation.

Low-complexity turbulence resilience enabled by a multi-mode bi-directional transceiver

Free-space optical (FSO) communication systems are acutely affected by the pointing issues and distortions that result from mechanical instability and environmental factors such as turbulence. These distortions have generally prevented single-mode bi-directional systems from being deployed without adaptive optics due to high optical losses. We investigate and compare the performance of both step and graded index multi-mode fibers for bi-directional communications over an emulated 400 m FSO channel. We propose that OM5 graded index fiber will simultaneously provide a near Gaussian optical transmission mode and a factor of greater than 5 increase in the field of view compared to single-mode fiber. We demonstrate that OM5 can support an error-free throughput of 10 Gbps for low-turbulence (D/r0 = 3) and 9.1 Gbps for high-turbulence (D/r0 = 9) using commercial bi-directional small form-factor pluggable (SFP+) transceivers with no adaptive optical components.

Urban Odyssey: “Pioneering multimodal routes for Tomorrow's smart cities”

Getting around in modern cities has become a daily puzzle for both residents and travelers. As cities increasingly evolve into complex hubs of innovation and development, the demand for efficient transportation solutions has never been higher. In the multifaceted field of urban transportation, cost-effective and efficient mobility remains a high priority. This paper looks at how cities are planning better ways for people to travel and move within the growing scope of multimodal transportation, a paradigm shift beyond traditional single-mode transit systems. Our approach to improving urban transport revolves around a few key principles: integration, innovation, and collaboration. Integration means bringing together different modes of transportation – like buses, trains, bikes, and even new technologies like ride-sharing services to make it easier for people to switch between them. This transformative approach not only aims to reduce congestion and reduce environmental footprints but also prioritizes user experience while ensuring a harmonious blend of convenience, sustainability, and accessibility. We use new technology and ideas to ensure that travel is easy, quick, and good for the environment. Looking at new trends and examples from around the world, this overview shows how cities are shaping a better future for everyone's daily commute. In this paper, we use Lingo software to solve a numerical example related to multi-modal transportation, demonstrating practical solutions for real-world implementation. Through innovation, we can find creative solutions to urban transportation challenges. Additionally, public transportation enhancements, such as the introduction of synchronized bus routes and electric vehicle charging stations, underline the commitment to sustainability and inclusivity. In the dynamic urban transportation landscape, cities will revolutionize our approach to providing cost-effective and efficient mobility solutions.

A Study of New Indexing Techniques for Multimodal Identification Using Iris, Fingerprint, and Face Biometrics

This paper explores innovative indexing techniques for multimodal biometric identification systems, focusing on iris, fingerprint, and facial recognition technologies. With the proliferation of digital identity verification needs, traditional single-modal biometric systems often fall short in terms of accuracy, speed, and security. Multimodal biometric systems, which integrate multiple biological characteristics, are becoming crucial for enhancing identification performance. This study aims to present a comprehensive analysis of new indexing methods that can significantly improve the efficiency, accuracy, and security of multimodal biometric identification systems.

Label-Free Continuous Cell Sorting Using Optofluidic Chip.

In the field of biomedicine, efficiently and non-invasively isolating target cells has always been one of the core challenges. Optical fiber tweezers offer precise and non-invasive manipulation of cells within a medium and can be easily integrated with microfluidic systems. Therefore, this paper investigated the mechanism of cell manipulation using scattering force with optical fiber tweezers. We employed flat-ended single-mode fiber to drive and sort cells and derived the corresponding scattering force formula based on the T-matrix model. A single-mode optical tweezers system for cell sorting was developed, and an optofluidic experimental platform was constructed that effectively integrates the optical system with microfluidic chips. The chip, featuring an expanded cross-channel design, successfully achieved continuous separation of yeast cells (8~10 µm in diameter) and polystyrene microspheres (15~20 µm in diameter), with a sorting efficiency of up to 86% and maintaining viability in approximately 90% of the yeast cells. Compared to other sorting systems, this system does not require labeling and can achieve continuous sorting with cell viability at a lower cost of instrumentation.

Automatic Lyric Transcription and Automatic Music Transcription from Multimodal Singing

Automatic lyric transcription (ALT) refers to transcribing singing voices into lyrics, while automatic music transcription (AMT) refers to transcribing singing voices into note events, i.e., musical MIDI notes. Despite these two tasks having significant potential for practical application, they are still nascent. This is because the transcription of lyrics and note events solely from singing audio is notoriously difficult due to the presence of noise contamination, e.g., musical accompaniment, resulting in a degradation of both the intelligibility of sung lyrics and the recognizability of sung notes. To address this challenge, we propose a general framework for implementing multimodal ALT and AMT systems. Additionally, we curate the first multimodal singing dataset, comprising N20EMv1 and N20EMv2, which encompasses audio recordings and videos of lip movements, together with ground truth for lyrics and note events. For model construction, we propose adapting self-supervised learning models from the speech domain as acoustic encoders and visual encoders to alleviate the scarcity of labeled data. We also introduce a residual cross-attention mechanism to effectively integrate features from the audio and video modalities. Through extensive experiments, we demonstrate that our single-modal systems exhibit state-of-the-art performance on both ALT and AMT tasks. Subsequently, through single-modal experiments, we also explore the individual contributions of each modality to the multimodal system. Finally, we combine these and demonstrate the effectiveness of our proposed multimodal systems, particularly in terms of their noise robustness.

Empowering artificial intelligence-based multi-biometric image sensor for human identification

Artificial intelligence (AI) and sensor technology developments have sparked revolutionary shifts in a number of fields, including biometric Identification. In order to improve human identification processes, this research offers a novel method that integrates AI and many biometric image sensors. The accuracy, robustness, and susceptibility to spoofing assaults of conventional single-modal biometric systems are among their many drawbacks. To overcome these challenges, we introduce a secure multi-biometric system that relies on feature-level fusion to identify users. In the preprocessing step, fingerprint images undergo Min-Max normalization to mitigate variations in image quality. In order to extract high-level features from both raw Electrocardiogram (ECG) signals and Min-Max normalized fingerprint images, ResNet50, a deep convolutional neural network, is used. These extracted feature vectors are able to distinguish between the two modalities. We proposed boosted Xgboost as a classifier for authentication in the identification steps to improve performance. The proposed approach is simulated using Python. A comparison study for improved Xgboost is presented using measures for accuracy, precision-recall, and F1-Score. Across all comparative metrics, the technique achieves much better performance. According to experimental findings, the suggested multi-biometric systems are more effective, dependable, and robust than the existing multi-biometric authentication systems.

Driver body condition monitoring system based on human-computer interaction

Tired driving is still a significant cause of traffic accidents, often causing huge economic losses to society.Therefore, how to avoid driver fatigue has become a meaningful problem. The purpose of this paper is to review and summarize various research methods of driver fatigue monitoring system at home and abroad, so as to reduce the probability of traffic accidents caused by fatigue driving. The main content of this paper is to review the existing fatigue driving monitoring technologies such as electrocardiogram (ECG), electroencephalogram (EEG), HRV, eye tracking system, CNN, facial state analysis, emotion recognition, 5G and brain wave recognition, and infer the fatigue level of drivers by using these technologies to extract various clues that usually represent the level of human vigilance. Assess cognitive load and identify emotional states to monitor tired driving. By discussing the advantages and disadvantages of various fatigue monitoring systems at home and abroad, this paper proposes an innovative method, which transforms the previous single-mode monitoring system into a multi-mode monitoring system, integrates multiple biosensors, extracts relevant features from the biosensor data and processes the data reasonably in order to understand the drivers physiological state more comprehensively. This method is expected to improve the accuracy of fatigue monitoring system.

A computable quantification of multi-mode Gaussian coherence and complete monogamy relation

We propose a quantification CνGn of n-mode Gaussian coherence. The value of CνGn only depends on the covariance matrices and displacement vectors of continuous-variable states without any optimization procedures, and thus is easily calculated. For n=1, CνG1 is a proper Gaussian coherence measure of single-mode CV system. For n≥2, CνGn is invariant under any permutation of submodes, nonincreasing under any n-mode local incoherent Gaussian channels, vanishes at incoherent Gaussian states, and, satisfies the unification condition and the hierarchy condition that a multi-partite quantum correlation measure should obey. Thus CνGn is a multi-partite Gaussian correlation measure, which reveals, though the quantum coherence lives in single-partite systems, that the multi-mode coherence for continuous-variable systems can be regarded as a multipartite Gaussian correlation between modes, and such multi-partite Gaussian correlation is also a quantum resource. Moreover, we show that CνGn is completely monogamous as a multipartite Gaussian correlation measure. This means that the n-mode Gaussian coherence subjects to the principles of resource allocation. In addition, CνGn is an upper bound of the geometric-based single-mode Gaussian coherence measure CBu by the Bures distance at pure Gaussian states of mode ≤2 and can be used to detect coherence in any n-mode Gaussian states more efficiently.

Utilization of nicking properties of CRISPR-Cas12a effector for genome editing

The CRISPR-Cas nickase system for genome editing has attracted considerable attention owing to its safety, efficiency, and versatility. Although alternative effectors to Cas9 have the potential to expand the scope of genome editing, their application has not been optimized. Herein, we used an enhanced CRISPR-Cas12a nickase system to induce mutations by targeting genes in a human-derived cell line. The optimized CRISPR-Cas12a nickase system effectively introduced mutations into target genes under a specific directionality and distance between nickases. In particular, the single-mode Cas12a nickase system can induce the target-specific mutations with less DNA double-strand breaks. By inducing mutations in the Thymine-rich target genes in single- or dual-mode, Cas12a nickase compensates the limitations of Cas9 nickase and is expected to contribute to the development of future genome editing technologies.

Quasi-direct Quantum Molecular Dynamics: The Time-Dependent Adaptive Density-Guided Approach for Potential Energy Surface Construction.

We present a new quasi-direct quantum molecular dynamics computational method which offers a compromise between quantum dynamics using a precomputed potential energy surface (PES) and fully direct quantum dynamics. This method is termed the time-dependent adaptive density-guided approach (TD-ADGA) and is a method for constructing a PES on the fly during a dynamics simulation. This is achieved by acquisition of new single-point (SP) calculations and refitting of the PES, depending on the need of the dynamics. The TD-ADGA is a further development of the adaptive density-guided approach (ADGA) for PES construction where the placement of SPs is guided by the density of the nuclear wave function. In TD-ADGA, the ADGA framework has been integrated into the time propagation of the time-dependent nuclear wave function and we use the reduced one-mode density of this wave function to guide when and where new SPs are placed. The PES is thus extended or updated if the wave function moves into new areas or if a certain area becomes more important. Here, we derive equations for the reduced one-mode density for the time-dependent Hartree (TDH) method and for multiconfiguration time-dependent Hartree (MCTDH) methods, but the TD-ADGA can be used with any time-dependent wave function method as long as a density is available. The TD-ADGA method has been investigated on molecular systems containing single- and double-minimum potentials and on single-mode and multi-mode systems. We explore different approaches to handle the fact that the TD-ADGA involves a PES that changes during the computation and show how results can be obtained that are in very good agreement with results obtained by using an accurate reference PES. Dynamics with TD-ADGA is essentially a black box procedure, where only the initialization of the system and how to compute SPs must be provided. The TD-ADGA thus makes it easier to carry out quantum molecular dynamics and the quasi-direct framework opens up the possibility to compute quantum dynamics accurately for larger molecular systems.

Gait Intention Prediction Using a Lower-Limb Musculoskeletal Model and Long Short-Term Memory Neural Networks.

The prediction of gait motion intention is essential for achieving intuitive control of assistive devices and diagnosing gait disorders. To reduce the cost associated with using multimodal signals and signal processing, we proposed a novel method that integrates machine learning with musculoskeletal modelling techniques for the prediction of time-series joint angles, using only kinematic signals. Additionally, we hypothesised that a stacked long short-term memory (LSTM) neural network architecture can perform the task without relying on any ahead-of-motion features typically provided by electromyography signals. Optical cameras and inertial measurement unit (IMU) sensors were used to track level gait kinematics. Joint angles were modelled using the musculoskeletal model. The optimal LSTM architecture in fulfilling the prediction task was determined. Joint angle predictions were performed for joints on the sagittal plane, benefiting from joint angle modelling using signals from optical cameras and IMU sensors. Our proposed method predicted the upcoming joint angles in the prediction time of 10 ms, with an averaged root mean square error of 5.3° and a coefficient of determination of 0.81. Moreover, in support of our hypothesis, the recurrent stacked LSTM network demonstrated its ability to predict intended motion accurately and efficiently in gait, outperforming two other neural network architectures: a feedforward MLP and a hybrid LSTM-MLP. The method paves the way for the development of a cost-effective, single-modal control system for assistive devices in gait rehabilitation.

Single-mode waveguides for GRAVITY

The second generation Very Large Telescope Interferometer (VLTI) instrument GRAVITY is a two-field infrared interferometer operating in the K band between 1.97 and 2.43 µm with either the four 8 m or the four 1.8 m telescopes of the Very Large Telescope (VLT). Beams collected by the telescopes are corrected with adaptive optics systems and the fringes are stabilized with a fringe-tracking system. A metrology system allows the measurement of internal path lengths in order to achieve high-accuracy astrometry. High sensitivity and high interferometric accuracy are achieved thanks to (i) correction of the turbulent phase, (ii) the use of low-noise detectors, and (iii) the optimization of photometric and coherence throughput. Beam combination and most of the beam transport are performed with single-mode waveguides in vacuum and at low temperature. In this paper, we present the functions and performance achieved with weakly birefringent standard single-mode fiber systems in GRAVITY. Fibered differential delay lines (FDDLs) are used to dynamically compensate for up to 6 mm of delay between the science and reference targets. Fibered polarization rotators allow us to align polarizations in the instrument and make the single-mode beam combiner close to polarization neutral. The single-mode fiber system exhibits very low birefringence (less than 23°), very low attenuation (3.6–7 dB km−1 across the K band), and optimized differential dispersion (less than 2.04 µrad cm2 at zero extension of the FDDLs). As a consequence, the typical fringe contrast losses due to the single-mode fibers are 6% to 10% in the lowest-resolution mode and 5% in the medium- and high-resolution modes of the instrument for a photometric throughput of the fiber chain of the order of 90%. There is no equivalent of this fiber system to route and modally filter beams with delay and polarization control in any other K-band beamcombiner.

Nonlinear topological laser on the non-Hermitian Haldane model with higher-order corner states.

The non-Hermitian skin effect (NHSE) on the non-Hermitian Haldane model with gain and loss on the honeycomb lattice with the outline of a triangle is discussed. The NHSE only occurs on the edge of the lattice, transforming the edge modes into the higher-order corner modes. The NHSE can also occur on a lattice with only loss, which can be treated as a lattice with gain and loss as well as a global loss added to it. When the saturated gain is added to the three corner sites of the dissipative lattice, a single-mode laser system is obtained. When any one site is stimulated initially, the system will reach a saturated state depending on the distribution of the corner modes, and the stable laser light is emitted by sites at the corners.

A Hybrid Brain–Computer Interface Combining P300 Potentials and Emotion Patterns for Detecting Awareness in Patients With Disorders of Consciousness

In this study, a hybrid brain-computer interface (BCI) system combining P300 potential and emotion patterns was proposed to improve the performance of awareness detection. Two video clips were flashed randomly to evoke the P300 potential, while a laughing or crying video clip was used to induce the corresponding emotion pattern. The subjects were asked to concentrate on the laughing or crying video clip cued by the instruction and to count the flashes of the corresponding video clip. Two layers of classification were developed. In the first layer, P300 detection and emotion recognition were performed separately using two support vector machine (SVM) classifiers. Specifically, the activation, spatial and connection patterns were fused in emotion recognition. In the second layer, the SVM scores of P300 detection and emotion recognition were fed into another SVM classifier to determine which video clip the subjects responded to. Six healthy subjects and eight patients with disorders of consciousness (DOC) were involved in the command-following experiment. The results showed that the accuracy of the hybrid BCI system was better than those of the single-modality systems. Furthermore, three patients were able to perform tasks (66%-72%) using our hybrid BCI, which indicated their residual awareness and emotion-related abilities.

Efficient Integrated Multimode Amplifiers for Scalable Long-Haul SDM Transmission

Scaling up cable capacity in a power- and cost-efficient manner is a goal of spatial-division multiplexing (SDM) in long-haul optical communication systems. Integration is a key to scaling SDM beyond the parallel single-mode fiber paradigm. Multicore or multimode fiber systems promise higher integration, but native multicore or multimode amplifiers pose challenges in efficient coupling of pump light into the amplifier, and in control of mode-dependent gain (MDG). We propose an integrated multimode amplifier design for signals in six spatial modes (12 spatial and polarization modes) that relies on excitation of only four pump spatial modes at 980 nm. Simulations of an exemplary amplifier predict a relatively flat gain above 10 dB, noise figure below 5 dB, and MDG standard deviation below 0.1 dB over the C-band (1530–1565 nm), with a power-conversion efficiency close to 30%. Key design features include a graded-index multimode fiber amplifier with optimized ring erbium doping profile, length and pump mode powers; and a cascade of wavelength- and mode-selective couplers that efficiently couples four pump diodes to four pump modes, while passing signal modes with minimal loss. Our design approach, which uses fewer pump diodes per signal mode than parallel single-mode systems, may provide an option for efficient, economical scaling of SDM long-haul systems.

Vehicle Target Recognition Method Based on Visible and Infrared Image Fusion Using Bayesian Inference

The accuracy of single-mode optical imaging systems for vehicle target recognition is limited by external ambient illumination or imaging equipment resolution. In this paper, a vehicle target recognition method based on visible and infrared image fusion using Bayesian inference is proposed. Based on the significance area detection combined with Spectral Residual (SR) and EdgeBox algorithms, the target area is marked on the visible light image, and the maximum between-class variance method and GrabCut algorithm are used to segment vehicle targets in marked images. The open operation filter is employed to extract the vehicle target features in the infrared image, and the fusion result of the visible light image and the infrared image is obtained by introducing the Intersection Over Union (IOU), and as the parameter of Bayesian inference, the class and attribute of the parameter are defined and substituted into the established naive Bayesian classification model, and the probability of the class is calculated to determine whether the vehicle target is recognized. Experiments under different test conditions were carried out, and the experimental results were as follows: the accuracy of image target recognition reached 77% when the vehicle target was not occluded; when the vehicle target was partially occluded, the accuracy of image target recognition reached 74%; the results verified that the proposed method can recognize vehicle targets in different scenarios.