Pattern Noise Research Articles

Real-Time DC Series Arc Fault Detection Based on Noise Pattern Analysis in Photovoltaic System

This paper presents a method for detecting series arc faults based on noise pattern analysis in photovoltaic systems. The arc detection circuit is required to detect the series arc quickly and not falsely detect the system noise as arc noise. This paper proposes noise pattern analysis to distinguish between system noise and arc noise and prevent false detection. First, periodic feature analysis prevents false detection by periodic noise, such as inverter noise, using the difference in periodic characteristics. Second, zero-range density analysis detects series arc noise using the difference in fluctuation characteristics. An integrated arc fault detection algorithm comprising discrete wavelet transform decomposition, periodic feature analysis, zero-range density, and an iterative loop algorithm is developed. The proposed methods are verified using noise distinction experiments, wherein the injected signal and arc noise are distinguished with high classification precision using the proposed noise pattern analysis method. The developed arc detection algorithm is applied to a real-time series arc detection board based on the TMS320F28335 DSP, and the performance of the series arc detection and false detection prevention is verified using the UL1699B arc detection test circuit and a real PV system.

Effects of Adding Monaural and Binaural Noise to a Dichotic Listening Task.

The Dichotic Digits Test (DDT) evaluates central auditory nervous system (CANS) dysfunction. The DDT is widely used in audiology clinics worldwide because it is clinically efficient and has good sensitivity and specificity for CANS lesions. However, the DDT shows a strong ceiling effect, which can mitigate its ability to detect subtle CANS dysfunction. This study examines the effects of adding monaural and binaural speech-spectrum noise to the DDT in an effort to make the test more taxing to the CANS and thereby reduce the observed ceiling effect. This was an experimental repeated measures study. The participants were 20 adults aged 18-50 years with normal, bilaterally symmetric peripheral hearing sensitivity. Each participant was administered one standard DDT test list (no noise added) and DDT test lists with binaural, monaural right, and monaural left noise added. For each of the noise-added conditions, lists were administered at two different signal-to-noise ratios (SNRs), for a grand total of seven DDT test lists per participant, presented in randomized order. Monaural and binaural noise effects on DDT scoring indices (Right and Left Ear Percent Correct Scores, Combined Total Percent Correct Scores, and Dichotic Difference Scores), as well as noise effects on the Right Ear Advantage (REA) for speech, were examined. Mixed model analyses of variance (ANOVAs) were used to examine fixed effects and interactions of Noise Condition and Ear. Adding noise to the standard DDT systematically reduced Right and Left Ear Percent Correct Scores and Combined Total Percent Correct Scores. Statistically significant differences on all indices were found between monaural and binaural noise-added conditions, suggesting a possible advantage for binaural listening in noise. These findings suggest that adding noise to tests of dichotic listening increases the difficulty of the task, and that further investigation of dichotic listening patterns in noise could potentially lead to more sensitive clinical evaluations of CANS integrity and function.

Multi-Task Learning for UAV Aerial Object Detection in Foggy Weather Condition

Adverse weather conditions such as haze and snowfall can degrade the quality of captured images and affect performance of drone detection. Therefore, it is challenging to locate and identify targets in adverse weather scenarios. In this paper, a novel model called Object Detection in a Foggy Condition with YOLO (ODFC-YOLO) is proposed, which performs image dehazing and object detection jointly by multi-task learning approach. Our model consists of a detection subnet and a dehazing subnet, which can be trained end-to-end to optimize both tasks. Specifically, we propose a Cross-Stage Partial Fusion Decoder (CSP-Decoder) in the dehazing subnet to recover clean features of encoder from complex weather conditions, thereby reducing the feature discrepancy between hazy and clean images, thus enhancing the feature consistency between different tasks. Additionally, to increase the feature modeling and representation capabilities of our network, we also propose an efficient Global Context Enhanced Extraction (GCEE) module to extract beneficial information from blurred images by constructing global feature context long-range dependencies. Furthermore, we propose a Correlation-Aware Aggregated Loss (CAALoss) to average noise patterns and tune gradient magnitudes across different tasks, accordingly implicitly enhancing data diversity and alleviating representation bias. Finally, we verify the advantages of our proposed model on both synthetic and real-world foggy datasets, and our ODFC-YOLO achieves the highest mAP on all datasets while achieving 36 FPS real-time detection speed.

PERFORMANCE OF 4253HT SMOOTHER BY DIFFERENT HANNINGS: APPLICATION IN RAINFALL DATA

Smoothing is an exploratory data analysis approach that focuses on removing noise or unstructured pattern from data series. This study mainly aims to compare the performance of 4253HT smoother in three types of Hannings and its application in forecasting. A sinusoidal signal was used where five different levels of contaminated normal noise were applied. Overall, 4253HT smoother with Shitan and Vazifean’s Hanning performs excellently over different percentages of noise, good at preserving edges, and able to travel closely with the signal of original pattern. The smoothed rainfall data gives a lower value of RMSE than the raw data which is 12.85 and 24.25 respectively. This concludes that the trend line obtained using smoothed data is more appropriate and reliable for forecasting. These results will be useful in predicting any time series data.

A Fast Approach to Optimize Tread Pattern Shape for Tire Noise Reduction

Impact force induced by tread pattern is one of the major mechanisms of tire noise; thus, how to reduce tire impact noise has been an important issue in regard to tire technology. In this paper, the mechanism of tire–pavement interaction noise is briefly described, and a prediction model for pattern noise is proposed. The prediction model was validated with indoor semi-anechoic chamber and pass-by noise experiments. Then, an optimization method for the tread pattern was proposed by using the basis vector method, and the synthetic pattern shape was generated through a linear combination of basis shape vectors. Finally, a novel multi-objective function was proposed, aimed at minimizing the impact noise generated by the tire pattern, and the weight factors of the basis vectors were optimized with a genetic algorithm. The method proposed in this paper can be used to evaluate or improve pattern performance and reduce trial and error in the pattern design stage.

A Novel Technique for Restoration of Images Subjected to Variable Noise Distribution

Image restoration plays a pivotal role in numerous applications across diverse fields due to its profound significance in enhancing the quality and interpretability of images. Whether in medical diagnostics, scientific research, surveillance, or historical preservation, the importance of image restoration cannot be overstated. By addressing issues such as noise, blur, and other forms of degradation, this process contributes to the preservation of vital information within images. This research focuses on enhancing image-denoising techniques by employing the Ladner-Fischer Adder (LFA) method. The study explores the effectiveness of LFA in mitigating varying noise distributions, particularly Salt and Pepper Noise (SPN) and Gaussian noise. Through detailed experimentation and analysis, the paper evaluates the performance of LFA-based Median Filtering (LFA-MF) and Finite Impulse Response (LFA-FIR) filtering methods. The results demonstrate substantial improvements in Peak Signal-to-Noise Ratio (PSNR), Mean Square Error (MSE), and Structural Similarity Index Measure (SSIM) values for both noise types. Notably, LFA-FIR consistently outperforms LFA-MF in terms of denoising efficacy, highlighting its potential for image restoration in the presence of complex noise patterns. Moreover, the research investigates the ASIC synthesis of LFA, examining parameters such as area, power consumption, and delay across different technology nodes. The findings emphasize the potential of LFA for robust image denoising under diverse noise conditions, offering insights into its applicability and performance in real-world scenarios.

Noise of a shrouded propeller due to ingestion of grid-generated turbulence

An experimental investigation of the noise produced by a shrouded propeller interacting with grid-generated turbulence in an anechoic wind tunnel is presented. Inflow turbulence statistics and the far-field noise directivity were measured. Measured inflow turbulence statistics and propeller operating parameters were used in a theoretical model to predict the turbulent inflow noise spectra and directivity. The theoretical model uses rapid distortion theory to calculate the distorted turbulent field at the propeller plane. The unsteady loading on the rotating blades is calculated using isolated blade response functions. The noise radiated to the far-field is then evaluated using a tailored Green's function approach which accounts for the acoustic scattering effect of the shroud. The theoretical model is able to capture the shape of the directivity pattern of the turbulent inflow noise produced by the shrouded propeller. The agreement between the measured and predicted spectra is reasonable, however the high frequency broadband levels tend to be underpredicted.

Initial experiences with Direct Imaging of Neuronal Activity (DIANA) in humans

Abstract Functional MRI (fMRI) has been widely used to study activity patterns in the human brain. It infers neuronal activity from the associated hemodynamic response, which fundamentally limits its spatiotemporal specificity. In mice, the Direct Imaging of Neuronal Activity (DIANA) method revealed MRI signals that correlated with extracellular electric activity, showing high spatiotemporal specificity. In this work, we attempted DIANA in humans. Five experimental paradigms were tested, exploring different stimulus types (flickering noise patterns, and naturalistic images), stimulus durations (50–200 ms), and imaging resolution (2 × 2 × 5 mm3 and 1 × 1 × 5 mm3). Regions of interest (ROI) were derived from Blood Oxygen Level Dependent (BOLD) fMRI acquisitions (both EPI and FLASH based) and T1-weighted anatomical scans. In Paradigm I (n = 1), using flickering noise patterns, signals were detected that resembled possible functional activity from a small ROI. However, changes in stimulus duration did not lead to corresponding signal changes (Paradigm II; n = 1). Therefore, care should be taken not to mistake artifacts for neuronal activity. In Paradigm III (n = 3), when averaged across multiple subjects, a ~200 ms long 0.02% signal increase was observed ~100 ms after the stimulus onset (10x smaller than the expected signal). However, white matter control ROIs showed similarly large signal fluctuations. In Paradigm IV (n = 3), naturalistic image stimuli were used, but did not reveal signs of a potential functional signal. To reduce partial voluming effects and improve ROI definition, in Paradigm V (n = 3), we acquired data with higher resolution (1 × 1 × 5 mm3) using naturalistic images. However, no sign of activation was found. It is important to note that repetitive experiments with short interstimulus intervals were found to be strenuous for the subjects, which likely impacted data quality. To obtain better data, improvements in sequence and stimulus designs are needed to maximize the DIANA signal and minimize confounds. However, without a clear understanding of DIANA’s biophysical underpinnings it is difficult to do so. Therefore, it may be more effective to first investigate DIANA signals with simultaneously recorded electrophysiological signals in more controlled settings, e.g., in anesthetized mice.

Deep Learning-Based Versus Iterative Image Reconstruction for Unenhanced Brain CT: A Quantitative Comparison of Image Quality.

This exploratory retrospective study aims to quantitatively compare the image quality of unenhanced brain computed tomography (CT) reconstructed with an iterative (AIDR-3D) and a deep learning-based (AiCE) reconstruction algorithm. After a preliminary phantom study, AIDR-3D and AiCE reconstructions (0.5 mm thickness) of 100 consecutive brain CTs acquired in the emergency setting on the same 320-detector row CT scanner were retrospectively analyzed, calculating image noise reduction attributable to the AiCE algorithm, artifact indexes in the posterior cranial fossa, and contrast-to-noise ratios (CNRs) at the cortical and thalamic levels. In the phantom study, the spatial resolution of the two datasets proved to be comparable; conversely, AIDR-3D reconstructions showed a broader noise pattern. In the human study, median image noise was lower with AiCE compared to AIDR-3D (4.7 vs. 5.3, p < 0.001, median 19.6% noise reduction), whereas AIDR-3D yielded a lower artifact index than AiCE (7.5 vs. 8.4, p < 0.001). AiCE also showed higher median CNRs at the cortical (2.5 vs. 1.8, p < 0.001) and thalamic levels (2.8 vs. 1.7, p < 0.001). These results highlight how image quality improvements granted by deep learning-based (AiCE) and iterative (AIDR-3D) image reconstruction algorithms vary according to different brain areas.

Analysis of Regional Ambient Seismic Noise in the Chukchi Sea Area in the Arctic Based on OBS Data from the Ninth Chinese National Arctic Scientific Survey

Ambient noise plays a crucial role in influencing the observation quality at seismic stations. By studying the distribution patterns of ambient noise, we can gain initial insights into the noise conditions within a specific research area. This paper investigates the properties of ambient noise in different frequency bands under environmental settings in the Chukchi Sea region, utilizing data collected from ocean bottom seismometers (OBSs) deployed during the Ninth Chinese National Arctic Scientific Survey. The probability density function (PDF) method is used to reveal the distinctive features of ambient noise. In addition, by comparing the crowed number values of ambient noise in the Chukchi Sea area with the global new low-noise model (NLNM) and new high-noise model (NHNM), a more comprehensive understanding of the patterns, distribution characteristics, and sources of ambient noise in the Arctic Chukchi Sea area is gained. The study suggests that the overlying sea ice in the Arctic Chukchi Sea area can suppress the microseismic band ambient noise, and the overall level of ambient noise in the Chukchi Sea area lies between the land seismic ambient noise level and the ambient noise level in the middle- and low-latitude sea areas. Meanwhile, an abnormal power spectrum caused by different levels of natural earthquakes is observed. This study fills the gap by using seafloor seismic instruments to investigate ambient noise in the Chukchi Sea area.

Thermal imaging spatial noise removal via deep image prior and step-variable total variation regularization

The imaging quality of an uncooled long wave infrared camera suffers severely from time-drifted spatial noises if not regulated by a thermoelectric cooler (TEC). These noises can be categorized within the frequency domain as either fixed pattern noise (FPN) or low-frequency noise (LPN). The time-drifted FPN is primarily governed by the temperature fluctuations of the focal plane array (FPA), whereas the LPN predominantly originates from non-uniform background radiations. Most existing methods have been proposed to address the FPN or LPN independently rather than from a unified perspective. In this paper, a self-supervised convolution neural network (CNN) is proposed to remove these two types of noise simultaneously. Because of the combination of the deep image prior (DIP) architecture and the total variation (TV) regularizer, this method is named TV-DIP. With the assumption that the spatial noises are constant in a short period of time, the spatial noise pattern is extracted directly from multiple continuous raw frames without ground-truth noise-free images. In this way, the noise pattern can be updated periodically during running time, without requiring the use of a mechanical shutter. Besides, the original TV regularizer is extended with a variable step, which shows a significant enhancement on the LPN denoising task. Furthermore, both simulation and experimental results have demonstrated that this approach achieves better performance in terms of the removal of spatial noise, the preservation of detail, and the suppression of artifacts. The source code and dataset are available on the website: https://github.com/brunolinux/TV-DIP

Estimation-free spatial-domain image reconstruction of structured illumination microscopy

Structured illumination microscopy (SIM) achieves super-resolution (SR) by modulating the high-frequency information of the sample into the passband of the optical system and subsequent image reconstruction. The traditional Wiener-filtering-based reconstruction algorithm operates in the Fourier domain, it requires prior knowledge of the sinusoidal illumination patterns which makes the time-consuming procedure of parameter estimation to raw datasets necessary, besides, the parameter estimation is sensitive to noise or aberration-induced pattern distortion which leads to reconstruction artifacts. Here, we propose a spatial-domain image reconstruction method that does not require parameter estimation but calculates patterns from raw datasets, and a reconstructed image can be obtained just by calculating the spatial covariance of differential calculated patterns and differential filtered datasets (the notch filtering operation is performed to the raw datasets for attenuating and compensating the optical transfer function (OTF)). Experiments on reconstructing raw datasets including nonbiological, biological, and simulated samples demonstrate that our method has SR capability, high reconstruction speed, and high robustness to aberration and noise.

Mapping of lattice distortion in martensitic steel—Comparison of different evaluation methods of EBSD patterns

To visualize the varying tetragonal distortions in high carbon martensitic steels by EBSD, two different approaches have been applied on backscattered Kikuchi diffraction (BKD) patterns. A band-edge refinement technique called Refined Accuracy (RA) (Oxford Instruments) is compared with a technique called Pattern Matching (PM), which optimizes the fit to a simulated BKD signal. RA distinguishes between hypothetical phases of different fixed c/a, while PM determines a best fitting continuous c/a by projective transformation of a master pattern. Both techniques require stored BKD patterns.The sensitivity of the c/a-determination was tested by investigating the microstructure of a ferritic steel with an expected c/a=1. The influence of the Kikuchi pattern noise on c/a was compared for a single or 40 averaged frames per measuring point, and turned out to be not significant. The application of RA and PM on the martensitic microstructure delivered qualitatively similar maps of c/a. The comparison of RA and PM shows that RA is suitably fast and precise during mapping the martensite c/a ratio in analyses of high carbon martensite, especially for fast initial surveys. As RA leads quantitatively to higher noise in c/a, the PM analysis can be used for higher precision results.

Relating clustered noise data to hospital patient satisfaction.

Hospital noise can be problematic for both patients and staff and consistently is rated poorly on national patient satisfaction surveys. A surge of research in the last two decades highlights the challenges of healthcare acoustic environments. However, existing research commonly relies on conventional noise metrics such as equivalent sound pressure level, which may be insufficient to fully characterize the fluctuating and complex nature of the hospital acoustic environments experienced by occupants. In this study, unsupervised machine learning clustering techniques were used to extract patterns of activity in noise and the relationship to patient perception. Specifically, nine patient rooms in three adult inpatient hospital units were acoustically measured for 24 h and unsupervised machine learning clustering techniques were applied to provide a more detailed statistical analysis of the acoustic environment. Validation results of five different clustering models found two clusters, labeled active and non-active, using k-means. Additional insight from this analysis includes the ability to calculate how often a room is active or non-active during the measurement period. While conventional LAeq was not significantly related to patient perception, novel metrics calculated from clustered data were significant. Specifically, lower patient satisfaction was correlated with higher Active Sound Levels, higher Total Percent Active, and lower Percent Quiet at Night metrics. Overall, applying statistical clustering to the hospital acoustic environment offers new insights into how patterns of background noise over time are relevant to occupant perception.

End-to-end Procedural Material Capture with Proxy-Free Mixed-Integer Optimization

Node-graph-based procedural materials are vital to 3D content creation within the computer graphics industry. Leveraging the expressive representation of procedural materials, artists can effortlessly generate diverse appearances by altering the graph structure or node parameters. However, manually reproducing a specific appearance is a challenging task that demands extensive domain knowledge and labor. Previous research has sought to automate this process by converting artist-created material graphs into differentiable programs and optimizing node parameters against a photographed material appearance using gradient descent. These methods involve implementing differentiable filter nodes [Shi et al. 2020] and training differentiable neural proxies for generator nodes to optimize continuous and discrete node parameters [Hu et al. 2022a] jointly. Nevertheless, Neural Proxies exhibits critical limitations, such as long training times, inaccuracies, fixed resolutions, and confined parameter ranges, which hinder their scalability towards the broad spectrum of production-grade material graphs. These constraints fundamentally stem from the absence of faithful and efficient implementations of generic noise and pattern generator nodes, both differentiable and non-differentiable. Such deficiency prevents the direct optimization of continuous and discrete generator node parameters without relying on surrogate models. We present Diffmat v2 , an improved differentiable procedural material library, along with a fully-automated, end-to-end procedural material capture framework that combines gradient-based optimization and gradient-free parameter search to match existing production-grade procedural materials against user-taken flash photos. Diffmat v2 expands the range of differentiable material graph nodes in Diffmat [Shi et al. 2020] by adding generic noise/pattern generator nodes and user-customizable per-pixel filter nodes. This allows for the complete translation and optimization of procedural materials across various categories without the need for external proprietary tools or pre-cached noise patterns. Consequently, our method can capture a considerably broader array of materials, encompassing those with highly regular or stochastic geometries. We demonstrate that our end-to-end approach yields a closer match to the target than MATch [Shi et al. 2020] and Neural Proxies [Hu et al. 2022a] when starting from initially unmatched continuous and discrete parameters.

DEGAN: Decompose-Enhance-GAN Network for Simultaneous Low-Light Image Lightening and Denoising

Images taken in low-light situations frequently have a significant quality reduction. Taking care of these degradation problems in low-light images is essential for raising their visual quality and enhancing high-level visual task performance. However, because of the inherent information loss in dark images, conventional Retinex-based approaches for low-light image enhancement frequently fail to accomplish real denoising. This research introduces DEGANet, a revolutionary deep-learning framework created particularly for improving and denoising low-light images. To overcome these restrictions, DEGANet makes use of the strength of a Generative Adversarial Network (GAN). The Decom-Net, Enhance-Net, and an Adversarial Generative Network (GAN) are three linked subnets that make up our novel Retinex-based DEGANet architecture. The Decom-Net is in charge of separating the reflectance and illumination components from the input low-light image. This decomposition enables Enhance-Net to effectively enhance the illumination component, thereby improving the overall image quality. Due to the complicated noise patterns, fluctuating intensities, and intrinsic information loss in low-light images, denoising them presents a significant challenge. By incorporating a GAN into our architecture, DEGANet is able to effectively denoise and smooth the enhanced image as well as retrieve the original data and fill in the gaps, producing an output that is aesthetically beautiful while maintaining key features. Through a comprehensive set of studies, we demonstrate that DEGANet exceeds current state-of-the-art methods in both terms of image enhancement and denoising quality.

Investigation on aerodynamic noise for leading edge erosion of wind turbine blade

Wind turbine blade is extremely prone to leading edge erosion under complicated operation conditions. This decreases the aerodynamic performance and power generation efficiency of wind turbine. Currently acoustic-based erosion detection of the blade still faces a significant challenge. Consequently, this paper presents aerodynamic noise mechanism of the eroded blade using the computational fluid dynamics method. The Zonal Detached Delay Eddy Simulation turbulence model is applied to obtain the three-dimensional instantaneous turbulent flow field. Then the far-field aerodynamic noise is carried out using the Ffowcs Williams and Hawkings approach. Simulations are performed for the erosion condition that determined by the length and depth at leading edge of 5-MW National Renewable Laboratory wind turbine. Results indicate that erosion leads to unsteady pressure pulsation near the leading edge, decrease of surface pressure difference in tip area, increase of airflow separation region and forward movement of separation point. Furthermore, the eroded blade achieves higher acoustic level than that of the normal one. From the noise directivity of wind turbine blade, erosion causes obvious asymmetry of noise pattern in the upwind and downwind directions.

Unsupervised Deep Embedded Fusion Representation of Single-Cell Transcriptomics

Cell clustering is a critical step in analyzing single-cell RNA sequencing (scRNA-seq) data, which allows us to characterize the cellular heterogeneity of transcriptional profiling at the single-cell level. Single-cell deep embedded representation models have recently become popular since they can learn feature representation and clustering simultaneously. However, the model still suffers from a variety of significant challenges, including the massive amount of data, pervasive dropout events, and complicated noise patterns in transcriptional profiling. Here, we propose a Single-Cell Deep Embedding Fusion Representation (scDEFR) model, which develop a deep embedded fusion representation to learn fused heterogeneous latent embedding that contains both the transcriptome gene-level information and the cell topology information. We first fuse them layer by layer to obtain compressed representations of intercellular relationships and transcriptome information. After that, the zero-inflated negative binomial model (ZINB)-based decoder is proposed to capture the global probabilistic structure of the data and reconstruct the final gene expression information and cell graph. Finally, by simultaneously integrating the clustering loss, crossentropy loss, ZINB loss, and the cell graph reconstruction loss, scDEFR can optimize clustering performance and learn the latent representation in fused information under a joint mutual supervised strategy. We conducted extensive and comprehensive experiments on 15 single-cell RNA-seq datasets from different sequencing platforms to demonstrate the superiority of scDEFR over a variety of state-of-the-art methods.

SiamNet: Exploiting source camera noise discrepancies using Siamese Network for Deepfake Detection

Recent advancements in deep neural networks especially GAN (Generative Adversarial Network) have resulted in the creation of more realistic deepfake media. This technology can swap the source person's face or alter facial expressions in an image or video; Media manipulated in such a way is termed deepfake. This type of manipulated media poses potential risks to journalism, politics, court proceedings and various social aspects. While existing approaches concentrate on the use of deep neural networks to directly extract facial artefacts for deepfake detection, they do not examine subtle inconsistencies in/across frame/frames. Moreover, state-of-the-art deepfake detection networks appear more complex and tend to overfit on specific artefacts which limits their generalizability on unseen data. This paper proposed a novel technique that tackles the problem of manipulated face detection in videos and images by exploiting the noise pattern inconsistency amongst face region and rest of the frame. To enable a comparison between the noise patterns of these two regions, we propose a two-stream Siamese-like network called SiamNet. This network can extract the noise patterns of the face region and patch through separate streams without increasing the number of parameters, thereby enhancing its efficiency and effectiveness. Each branch consists of pretrained Inception-v3 architecture for camera noise extraction. Siamese training is utilized to compare both noise patterns computed through different base models. The proposed two-branch network, SiamNet is found efficient for several large-scale deepfake datasets such as FF++, Celeb-DF, DFD and DFDC achieving accuracy rates of 99.7%, 98.3%, 96.08% and 89.2% respectively. Furthermore, the proposed technique exhibits greater generalizability and outperforms state-of-the-art of deepfake detection methods. Performance of the proposed model is also evaluated on FaceForensics benchmark dataset against different approaches.

Bidirectionally Photoresponsive Optoelectronic Transistors with Dual Photogates for All‐Optical‐Configured Neuromorphic Vision

AbstractBio‐inspired neuromorphic vision sensors, integrating optical sensing, and processing functions have attracted significant attention for developing future low‐power and high‐efficiency imaging systems. However, the compulsory electrical signal modulation to achieve inhibitory behaviors in most reported neuromorphic vision sensors results in additional hardware and computational latency. Herein, bidirectional photoresponsive optoelectronic synapses based on In2O3/Al2O3/Y6 phototransistors are achieved, realizing all‐optical‐configured synaptic weight updates enabled by dual photogates. The inhibitory and excitatory photoresponses originate from the photogating effects provided by trapped photogenerated electrons in Al2O3 under near‐infrared light and the ionized oxygen vacancies in In2O3 under ultra‐violet light, respectively. The bidirectional phototransistor illustrates outstanding optoelectronic synaptic characteristics with low nonlinearity and asymmetry, demonstrating high efficiencies in both preprocessing and postprocessing tasks, such as noise reduction, contrast enhancement, and pattern recognition. The proposed dual‐photogate optoelectronic synapses provide effective strategies to construct high‐efficiency neuromorphic vision sensors and in‐sensor computing systems.