Recognition Accuracy Rate Research Articles

Информационно-аналитическая система детектирования движения объектов на пешеходном переходе

To manage pedestrian and vehicle flows at intersections, systems are implemented that use adaptive traffic light models, adjusting time intervals based on the volume of pedestrians and vehicles. These systems include video cameras that monitor road user movements, enhancing real-time traffic control. This paper introduces an information and analytical system for managing transport and pedestrian flows using the YOLO neural model, which enables object recognition. The system performs several operations: converting the original image to gray scale, applying Gaussian blurring, detecting object boundaries through the Canny filter and fuzzy logic for edge detection, and contour processing, where each identified contour is assigned a unique number. The neural network then compares detected contours to the training sample data, determining whether the object is a person or a vehicle. The paper presents experimental results for these algorithms in object recognition. Modified software and images of intersections with pedestrian crossings captured from street video cameras in Kursk were used in the experiments. The recognition accuracy rate from the experiments was 72.4%.

User emotion recognition and indoor space interaction design: a CNN model optimized by multimodal weighted networks

In interior interaction design, achieving intelligent user-interior interaction is contingent upon understanding the user’s emotional responses. Precise identification of the user’s visual emotions holds paramount importance. Current visual emotion recognition methods rely solely on singular features, predominantly facial expressions, resulting in inadequate coverage of visual characteristics and low recognition rates. This study introduces a deep learning-based multimodal weighting network model to address this challenge. The model initiates with a convolutional attention module, employing a self-attention mechanism within a convolutional neural network (CNN). As a result, the multimodal weighting network model is integrated to optimize weights during training. Finally, a weight network classifier is derived from these optimized weights to facilitate visual emotion recognition. Experimental outcomes reveal a 77.057% correctness rate and a 74.75% accuracy rate in visual emotion recognition. Comparative analysis against existing models demonstrates the superiority of the multimodal weight network model, showcasing its potential to enhance human-centric and intelligent indoor interaction design.

Signal recognition and prediction system based on random forest model

To accurately predict the occurrence of rock bursts during deep coal mining and ensure the safe and efficient operation of coal mines, a signal recognition and prediction system based on the Random Forest model is proposed. This system utilizes electromagnetic radiation (EMR) and acoustic emission (AE) signal data, employing feature extraction and point biserial correlation analysis to screen out the most relevant feature parameters. A Random Forest binary classification model is constructed to identify interference signals. Subsequently, by introducing new features such as moving average slope and exponentially weighted moving average (EWMA), a time series analysis of precursor feature signals is conducted. A real-time warning model based on the Random Forest algorithm is developed, dynamically calculating the probability of precursor feature signal occurrence by integrating historical data and real-time data changes. This approach improves the accuracy and recall rate of signal recognition and prediction, providing reliable data support for mine safety management.

Room-temperature self-calibrating sensor based on CsPbBr3/SnO2 for detecting low-concentration sulfur dioxide

Sulfur dioxide (SO2) is a colorless toxic gas, and accurate monitoring of its concentration is important for safety. However, accuracy is often interfered by humidity. To address the issues, in this work, a self-calibrating SO2 sensor based on CsPbBr3/SnO2 was prepared, the sensor exhibited an excellent response to parts per billion (ppb) levels of SO2 at room temperature, which is rarely achieved by other SO2 sensors. The response sensitivity of the sensor to 600 ppb SO2 reached up to 0.53 with response/recovery times of 17/285 s. In addition, the CsPbBr3/SnO2-based SO2 gas sensor showed a low detection limit, good reproducibility, and high selectivity. Moreover, theoretical calculations show that the gas recognition mechanism is attributed to the significant changes in the band structure and density of states of the heterojunction materials CsPbBr3/SnO2 after gas adsorption, thereby altering the charge carrier transport behavior and electrical properties. A self-calibrating algorithm MobileNetV2 was employed to eliminate humidity interference and effectively identify low-concentration SO2 in humid environments, and the recognition accuracy rate is 0.85. This work demonstrates the production of a high-performance SO2 gas sensor, and provides an effective method to eliminate the interference of humidity, making it significant in environmental monitoring and air quality control.

Power text information extraction based on multi-task learning

In order to improve the analysis and processing speed of power system fault text in actual business scenarios, a power fault text information automatic extraction model based on pre-training and multi-task learning is proposed. The pre-training model is used to learn the context information of power text words, and the first-order and second-order fusion features of words are mined to enhance the representation ability of features. The multi-task learning framework is used to combine the learning of named entity recognition and relationship extraction tasks to achieve mutual complementation and mutual promotion between entity recognition and relationship extraction, thereby improving the performance of power fault text information extraction. Finally, the model is verified by the daily business data of a power grid data center. Compared with other models, the accuracy and recall rate of power fault text entity recognition and relationship extraction are improved.

Class‐wised domain decoupling‐guided adversarial feature learning for cross‐age face recognition

AbstractHow to extract age‐independent identity features from face images has long been the major challenge in cross‐age face recognition task. In this letter, a class‐wised domain decoupling‐guided adversarial feature learning network to extract age‐independent face features is proposed. In this model, an adversarial feature learning module is designed to decompose identity features and age features. Considering that artificially designed age interference suppression strategies are difficult to fit the complex relationship between identity features and age features, an adversarial training strategy incorporated with an attention mechanism in the adversarial feature learning module is introduced to suppress age interference adaptively. Besides, under the idea of maximizing the distribution difference of identity feature domain and age feature domain, a class‐wised domain decoupling module is further designed to guide the model to extract age‐independent identity features. The proposed model has been validated on the well‐known public datasets CALFW and CACD‐VS, where it achieved remarkable recognition accuracy rates of 94.5% and 99.5%, respectively. This represents improvements of 4.2% and 0.1% over the state‐of‐the‐art single‐task‐based models, clearly showcasing the effectiveness of the proposed class‐wised domain decoupling‐guided adversarial feature learning model.

Speech recognition in videos using features of different enhancement filters

Video Speech Recognition (VSR) is the ability of software to extract spoken text from input sources, such as offline videos and YouTube. It achieves a good recognition accuracy rate for videos with clean environments. However, with unwanted noisy elements in speech, this presents a major challenge. Depending on where the noise originates from, there are considerable differences in the impact of various noise kinds on video speech recognition, including background music, multiple speakers, microphone quality, and dialects. It is not possible to create a filter for every kind of noise because the sources of noise might differ greatly and the speech environment’s circumstances can change over time. Hence, instead of using a single filter for each type of noise, this study proposed to combine features from a small number of noise removal filters applied to the same video speech signal, resulting in a better VSR output. According to experimental results, using the proposed framework increases the accuracy of video speech recognition when evaluated on YouTube videos with different noise levels. In test experiments, the proposed framework was evaluated against state-of-the-art approaches. The accuracy mean of the relevant current approaches was improved by 16.67% and against the best accuracy of them by 7.78%. This research contributes to the video speech recognition topic as the proposed framework can facilitate the extraction of spoken phrases in millions of hours of video available online, indexing them, and then searching for them through their index.

Artificial optoelectronic synapse based on CdSe nanobelt photosensitized MoS2 transistor with long retention time for neuromorphic application

Abstract Optoelectronic synaptic devices have been regarded as the key component in constructing neuromorphic computing systems. However, the optoelectronic synapses based on conventional 2D transistor are still suffering low photosensitivity and volatile retention behavior, which can affect the recognition accuracy and long-term memory. Here, a novel optoelectronic synaptic device based on surface-state-rich CdSe nanobelt photosensitized 2D MoS2 transistor is demonstrated. Benefiting from the excellent light absorption of CdSe and effective charge trapping at the hetero-interface, the device exhibits not only high photosensitivity but also long retention time (>1,500 s). In addition, typical synaptic functions including the excitatory postsynaptic current, paired-pulse facilitation, the transformation from short-term to long-term plasticity, the transformation from short-term to long-term plasticity, spike-amplitude-dependent plasticity, and learning-forgetting-relearning process are successfully simulated and modulated by light stimulation. Most importantly, an artificial neural network is simulated based on the optical potentiation and electrical habituation characteristics of the synaptic devices, with recognition accuracy rates of 89.2, 93.8, and 91.9 % for file type datasets, small digits, and large digits are achieved. This study demonstrates a simple and efficient way to fabricate highly photosensitive optoelectronic synapse for artificial neural networks by combining the merits of specific materials and device architecture.

Deep learning based assembly process action recognition and progress prediction facing human-centric intelligent manufacturing

With the breakthroughs in cutting-edge technologies and the growing demand for personalized services, the approach of full automation alone is no longer sufficient to meet rapidly changing production needs. In light of this, a new human-centric manufacturing (HCM) paradigm has been put forward. Industrial human action recognition, as one of the key means of realizing HCM, has received widespread attention and research in recent years. In this paper, an end-to-end deep learning-based action recognition and progress prediction model was built, which can not only recognize action classes in real-time, but also accurately predict action completion times, to improve the efficiency and safety of HCM systems. First, a non-contact data acquisition system based on Azure Kinect sensor, MediaPipe and You-Only-Look-Once V5 (Yolo V5) algorithm was constructed, achieving comprehensive perception of human and objects in manufacturing system. Second, an Auto-encoder model was established to learn and compress high-dimensional human joint and object data into low-dimensional space to achieve efficient representation of the operation process. Finally, the human action recognition and progress prediction model was built and evaluated in a reducer assembly line, with an accuracy rate for action recognition of 99 %, and a time error for progress prediction of less than 6.9 %. This research aims to provide guidance for improving the level of intelligence in manufacturing and the application of AI technology in the industrial field.

Rapid prediction and visualization of safe moisture content in alfalfa seeds based on multispectral imaging technology

Moisture significantly impacts seed sales, storage, and processing. Traditional moisture testing methods are often slow, labor-intensive, and inadequate for the rapid detection demands of modern agriculture, particularly for non-destructive testing of individual seeds. This study applied multispectral imaging to obtain morphological and spectral data from alfalfa seeds at six moisture levels (4 %, 8 %, 12 %, 16 %, 25 %, and 41 %). By integrating algorithms such as Support Vector Machines (SVM), Random Forests (RF), Linear Discriminant Analysis (LDA), Back Propagation Neural Network (BPNN), and normalized typical discriminant analysis (nCDA) algorithms, classification models were developed to distinguish between safe and unsafe moisture levels. The Results indicated that spectral data alone significantly improved model accuracy and prediction. nCDA visualizations effectively illustrated spatial moisture distribution, highlighting stark color differences between seeds in the safe moisture range (4 %, 8 %, 12 %) and those in the unsafe range (16 %, 25 %, 41 %). BPNN exhibited high model precision, achieving a recognition accuracy rate of 90.1 % for safe and unsafe moisture content. Key wavelengths identified by the Permutation method included 970, 880, 570, and 490 nm. Pearson correlation analysis showed a significant positive correlation between germination indicators and spectral data, which strengthened with longer seed storage. These findings confirm the potential of multispectral imaging for assessing the safe moisture content of alfalfa seeds, supporting the development of detection systems for evaluating moisture content in individual seeds. This advancement enables the rapid removal of high-moisture seeds, preventing deterioration during storage.

Study on automatic and rapid diagnosis of distal radius fracture by X-ray

This article aims to combine deep learning with image analysis technology and propose an effective classification method for distal radius fracture types. Firstly, an extended U-Net three-layer cascaded segmentation network was used to accurately segment the most important joint surface and non joint surface areas for identifying fractures. Then, the images of the joint surface area and non joint surface area separately were classified and trained to distinguish fractures. Finally, based on the classification results of the two images, the normal or ABC fracture classification results could be comprehensively determined. The accuracy rates of normal, A-type, B-type, and C-type fracture on the test set were 0.99, 0.92, 0.91, and 0.82, respectively. For orthopedic medical experts, the average recognition accuracy rates were 0.98, 0.90, 0.87, and 0.81, respectively. The proposed automatic recognition method is generally better than experts, and can be used for preliminary auxiliary diagnosis of distal radius fractures in scenarios without expert participation.

Self and mother referential processing in phonological false memory

ABSTRACT The current study used phonological DRM and self-reference paradigms to examine whether phonological false memory was affected by self and mother referential processing. In Experiment 1 and 2, words in the Deese–Roediger–McDermott lists were presented to participants with their own name, their mother’s name or “Lu Xun”. Words were presented with their own name and “Lu Xun” in Experiment 3 to compare the effect of self and other referential processing specifically. The results revealed that true recognition rates were not different between self-reference and mother-reference conditions and both were higher than the other-reference condition; there were no significant differences in false recognition rates among the self-reference, mother-reference and other-reference conditions. The findings are discussed in terms of theories of the self and mother reference effects and phonological false memory.

Skin-inspired multimodal tactile sensor aiming at smart space extravehicular multi-finger operations

Tactile sensing of skin shapes the interactions between hand and the surrounding world, owing to the remarkable natural sensory system. But for astronauts, tactile feedback cannot be obtained biologically due to the very thick protective gloves, which seriously hinders the flexibility of the extravehicular activity. In this work, inspired by human skin, we develop a biomimetic multi-layer tactile sensor (BMLTS) that can work like the fast adaptive and slow adaptive receptors of fingertip to achieve the multimodal tactile perception based on the fusion of thermosensitive, piezoelectric, triboelectric and piezoresistive materials. Based on the BMLTS, the tactile perception for temperature, surface roughness discrimination and smart object grasping is successfully achieved, which fully simulates the basic tactile perceptions in typical extravehicular operations. Furthermore, by the combination of BMLTS and deep learning, a biomimetic intelligent perception system (BIPS) that can make decisions to the tactile signal just like human brain is constructed, which can provide real-time tactile information for astronauts. Intelligent real-time object material identification, writing and recording are conducted using BIPS, reaching the recognition accuracy rate higher than 95%. This work lays the foundation for the systematization and integration of tactile sensors, and paves the way for the development of dexterous tactile perception for smart extravehicular activities of astronauts.

Recognition of domestic <em>Felis Catus</em> vocalisations using a computational associative model

The cat (Felis catus) lives around or with humans and is considered one of the most popular pets in the world. Domestic cats have an extensive and complex vocalisation repertoire, and little is known about the meaning of their vocalisations; these animal sounds have been studied to determine their behavioural characteristics. This work implemented a computational model for cat vocalisation as a signal process to characterise it. Additionally, a recognition model is proposed to identify and associate domestic cat vocalisations with cat intentions. Domestic cat vocalisations were studied by obtaining sounds from audio and video sources stored in the web cloud and processing signals (distress, hiss, howl, meow, fright, purr, sexual call, and trill vocalisations) that belong to some cat situations (soliciting attention, complaining, surprise, threatening, welcoming, fighting, feeling hungry, mating, etc.). In measured signals, environmental noise was removed using a computational algorithm to increase the effectiveness of the recognition process. As a first approach, cat vocalisation patterns were extracted using Welch's method for spectral density estimation, appreciating vocalisation frequency characteristics and statistics. Additionally, a similarity function to compare cat vocalisation signals was developed, excluding useless patterns that could introduce computational problems in the learning phase of the recognition model. Then, an associative memory, the Lernmatrix Thresholded model, was used to recognise the domestic cat vocalisation needed, and pair associations between sounds and moods related to the recognition process result were established, i.e., identifying cat intention through vocalisation by analysing the signals emitted. The experimental results show the performance of the thresholded Lernmatrix among different pattern recognition models, where accurate vocalisation recognition rates are obtained. This computational model emerges as a suitable tool that could be helpful in animal care and ethology science to characterise the behaviour of animals through their vocalisation.

Bayesian-optimized LSTM-DWT approach for reliable fault detection in MMC-based HVDC systems

As Europe integrates more renewable energy resources, notably offshore wind power, into its super meshed grid, the demand for reliable long-distance High Voltage Direct Current (HVDC) transmission systems has surged. This paper addresses the intricacies of HVDC systems built upon Modular Multi-Level Converters (MMCs), especially concerning the rapid rise of DC fault currents. We propose a novel fault identification and classification for DC transmission lines only by employing Long Short-Term Memory (LSTM) networks integrated with Discrete Wavelet Transform (DWT) for feature extraction. Our LSTM-based algorithm operates effectively under challenging environmental conditions, ensuring high fault resistance detection. A unique three-level relay system with multiple time windows (1 ms, 1.5 ms, and 2 ms) ensures accurate fault detection over large distances. Bayesian Optimization is employed for hyperparameter tuning, streamlining the model’s training process. The study shows that our proposed framework exhibits 100% resilience against external faults and disturbances, achieving an average recognition accuracy rate of 99.04% in diverse testing scenarios. Unlike traditional schemes that rely on multiple manual thresholds, our approach utilizes a single intelligently tuned model to detect faults up to 480 ohms, enhancing the efficiency and robustness of DC grid protection.

Two-Stage Detection Algorithm for Plum Leaf Disease and Severity Assessment Based on Deep Learning

Crop diseases significantly impact crop yields, and promoting specialized control of crop diseases is crucial for ensuring agricultural production stability. Disease identification primarily relies on human visual inspection, which is inefficient, inaccurate, and subjective. This study focused on the plum red spot (Polystigma rubrum), proposing a two-stage detection algorithm based on deep learning and assessing the severity of the disease through lesion coverage rate. The specific contributions are as follows: We utilized the object detection model YOLOv8 to strip leaves to eliminate the influence of complex backgrounds. We used an improved U-Net network to segment leaves and lesions. We combined Dice Loss with Focal Loss to address the poor training performance due to the pixel ratio imbalance between leaves and disease spots. For inconsistencies in the size and shape of leaves and lesions, we utilized ODConv and MSCA so that the model could focus on features at different scales. After verification, the accuracy rate of leaf recognition is 95.3%, and the mIoU, mPA, mPrecision, and mRecall of the leaf disease segmentation model are 90.93%, 95.21%, 95.17%, and 95.21%, respectively. This research provides an effective solution for the detection and severity assessment of plum leaf red spot disease under complex backgrounds.

Research on automated rapid diagnosis of distal radius fractures using X-ray images

The primary diagnostic method for distal radius fractures remains conventional X-ray imaging. In emergency situations, rapid assessment and decision-making are often required from non-orthopedic physicians or junior radiologists in the emergency room. Therefore, an accurate and efficient auxiliary diagnostic technology for distal radius fractures is of great significance. This paper combines deep learning with image analysis techniques to propose an effective classification method for types of distal radius fractures. Firstly, an extended U-Net three-level cascade segmentation network is used for precise segmentation of the most critical joint surface and non-joint surface areas for fracture identification. The edge is further optimized using the central pixel image block classification method. Then, fracture identification is performed separately on the joint surface and non-joint surface areas. Finally, based on the classification results of the two types of images, a comprehensive judgment is made to determine normal or ABC fracture types. The experimental X-ray data comes from three tertiary hospitals, with a total of 12,000 images in the training set, 3,000 of each type, and 1,200 images in the test set, 300 of each type. To further demonstrate the generalization performance of the classification experiment, an additional 500 images were added for fracture classification testing, coming from collaborative medical institutions in the United States and Germany. All images were annotated by orthopedic medical experts with more than ten years of experience. The accuracy rates for normal, type A, type B, and type C fractures in the test set were 0.99, 0.92, 0.91, and 0.82, respectively. For orthopedic medical experts, the average recognition accuracy rates were 0.98, 0.90, 0.87, and 0.81. The proposed automatic recognition method overall performs better than experts and can be used for preliminary auxiliary diagnosis of distal radius fractures without expert participation.

Research on automated rapid diagnosis of distal radius fractures using X-ray images

The primary diagnostic method for distal radius fractures remains conventional X-ray imaging. In emergency situations, rapid assessment and decision-making are often required from non-orthopedic physicians or junior radiologists in the emergency room. Therefore, an accurate and efficient auxiliary diagnostic technology for distal radius fractures is of great significance. This paper combines deep learning with image analysis techniques to propose an effective classification method for types of distal radius fractures. Firstly, an extended U-Net three-level cascade segmentation network is used for precise segmentation of the most critical joint surface and non-joint surface areas for fracture identification. The edge is further optimized using the central pixel image block classification method. Then, fracture identification is performed separately on the joint surface and non-joint surface areas. Finally, based on the classification results of the two types of images, a comprehensive judgment is made to determine normal or ABC fracture types. The experimental X-ray data comes from three tertiary hospitals, with a total of 12,000 images in the training set, 3,000 of each type, and 1,200 images in the test set, 300 of each type. To further demonstrate the generalization performance of the classification experiment, an additional 500 images were added for fracture classification testing, coming from collaborative medical institutions in the United States and Germany. All images were annotated by orthopedic medical experts with more than ten years of experience. The accuracy rates for normal, type A, type B, and type C fractures in the test set were 0.99, 0.92, 0.91, and 0.82, respectively. For orthopedic medical experts, the average recognition accuracy rates were 0.98, 0.90, 0.87, and 0.81. The proposed automatic recognition method overall performs better than experts and can be used for preliminary auxiliary diagnosis of distal radius fractures without expert participation.

Driverless holographic simulation test integrated environment design

Abstract Aiming at the design problem of simulation test integrated environment for unmanned car, starting from the four aspects of unmanned simulation test [1], real scene quiz, scene analysis and layout, and data visualization and processing, through the combination of software and hardware, the use of various types of sensors, controllers, communication equipment, etc. for collaborative management of unmanned car driving, and the use of intelligent networked cars to collaboratively verify the multi-class algorithms and driving programs, and continuously improve the Overall environment design [2] [3], the integrated environment design is getting better and better. This overall environment is tested based on carla’s Apollo autonomous driving simulation software, and the test results show that the recognition accuracy rate reaches 87% and the simulation rate reaches 91%, etc. It shows that the unmanned holographic simulation test integrated environment technology program is feasible.

Flexible arch-shaped triboelectric sensor based on 3D printing for badminton movement monitoring and intelligent recognition of technical movements

Efficient monitoring and recognition of movement are crucial in enhancing athletic performance. Traditional methods have limitations in terms of high site requirements and power consumption, making them unsuitable for long-term tracking and monitoring. A potential solution to low-power monitoring of body area networks is triboelectric sensors. However, the current analysis method for badminton triboelectric sensing data is relatively simple, while flexible, triboelectric sensors based on 3D printing face issues such as discomfort when joints are bent or twisted in a large range. In light of this, a flexible arch-shaped triboelectric sensor based on 3D printing (FA-Sensor) is proposed. By combining neural network algorithms with the signal acquisition module and the master computer, an intelligent multi-sensor node system for badminton monitoring is established. The FA-Sensor exhibits high sensitivity to bending and twisting motions due to its elastic TPE shell and arched shape design. It minimizes interference with human motion during bending (10°–150°) or twisting (20°–100°) over a wide range. The peak output voltage of the FA-Sensor demonstrates a clear functional relationship with the bending angle, exhibiting piecewise sensitivities of 7.98 and 29.28 mV/°, respectively. For seven different parts of the human body, it can be quickly customized to different sizes, with stable and repeatable response outputs. In application, the badminton sports monitoring system enables real-time feedback and recognition of four typical technical movements, achieving a recognition accuracy rate of 97.2%. The system enables athletes to analyze and enhance badminton technology while also exhibiting promising potential for application in other intelligent sports domains.