Final Recognition Research Articles

Deep learning via CNN for identification of blue core phenomenon in helicon plasma discharge

Based on deep learning image recognition techniques, a convolutional neural network model for discharge mode recognition of helicon plasma was trained. The accuracy of the model was evaluated using functions such as F1-scores and the confusion matrix. The final recognition accuracy was more than 98.18% after 30 iterations. Interpretable analysis was done using methods such as gradient-weighted class activation mapping to verify the model's robustness as well as repeatability. The model identification results were compared with Langmuir probe diagnostic results. It was found a good fit between the model and the probe results, corroborating the correctness of the model. The present model can well identify the critical power of entering W mode in the discharge process of helicon plasma. As the discharge database expands, it has great potential for recognizing the higher-order discharge modes based on deep learning.

A residual deep learning network for smartwatch-based user identification using activity patterns in daily living

User identification is a critical aspect of smartwatch security, ensuring that only authorized individuals gain access to sensitive information stored on the device. Conventional methods like passwords and biometrics have limitations, such as the risk of forgetting passwords or the potential for biometric data to be compromised. This research proposes a novel approach for user identification on smartwatches by analyzing activity patterns using a hybrid residual neural network called Att-ResBiLSTM. The proposed method leverages unique patterns of user interactions with their smartwatches, including application usage, typing behavior, and motion sensor data, to create an individualized user profile. Employing a deep learning network specifically designed for wearable devices, the system can reliably and promptly identify users by analyzing their activity patterns. The Att-ResBiLSTM architecture comprises three key components: convolutional layers, ResBiLSTM, and an attention layer. The convolutional layers extract spatial features from the pre-processed data. At the same time, the ResBiLSTM component captures long-term dependencies in the time-series data by combining the advantages of bidirectional long short-term memory (BiLSTM) and residual connections. The attention mechanism enhances the final recognition features by selectively prioritizing the most informative elements of the input data. The Att-ResBiLSTM model is trained and evaluated using a diverse dataset of user activity patterns. Experimental results demonstrate that the proposed approach achieves remarkable accuracy in user identification, with an accuracy rate of 98.29% and the highest F1-score of 98.24%. The research also conducts a comparative analysis to assess the efficacy of accelerometer data versus gyroscope data, revealing that combining both sensor modalities improves user identification performance. The proposed methodology provides a reliable and user-friendly alternative to conventional user authentication techniques for smartwatches. This approach leverages activity patterns and a hybrid residual deep learning network to offer a robust and efficient solution for user identification based on smartwatch data, thereby enhancing the overall security of wearable devices.

A Modified Transformer Network for Seizure Detection Using EEG Signals.

Seizures have a serious impact on the physical function and daily life of epileptic patients. The automated detection of seizures can assist clinicians in taking preventive measures for patients during the diagnosis process. The combination of deep learning (DL) model with convolutional neural network (CNN) and transformer network can effectively extract both local and global features, resulting in improved seizure detection performance. In this study, an enhanced transformer network named Inresformer is proposed for seizure detection, which is combined with Inception and Residual network extracting different scale features of electroencephalography (EEG) signals to enrich the feature representation. In addition, the improved transformer network replaces the existing Feedforward layers with two half-step Feedforward layers to enhance the nonlinear representation of the model. The proposed architecture utilizes discrete wavelet transform (DWT) to decompose the original EEG signals, and the three sub-bands are selected for signal reconstruction. Then, the Co-MixUp method is adopted to solve the problem of data imbalance, and the processed signals are sent to the Inresformer network for seizure information capture and recognition. Finally, discriminant fusion is performed on the results of three-scale EEG sub-signals to achieve final seizure recognition. The proposed network achieves the best accuracy of 100% on Bonn dataset and the average accuracy of 98.03%, sensitivity of 95.65%, and specificity of 98.57% on the long-term CHB-MIT dataset. Compared to the existing DL networks, the proposed method holds significant potential for clinical research and diagnosis applications with competitive performance.

A Highly Sensitive Strain Sensor with Self-Assembled MXene/Multi-Walled Carbon Nanotube Sliding Networks for Gesture Recognition.

Flexible electronics is pursuing a new generation of electronic skin and human-computer interaction. However, effectively detecting large dynamic ranges and highly sensitive human movements remains a challenge. In this study, flexible strain sensors with a self-assembled PDMS/MXene/MWCNT structure are fabricated, in which MXene particles are wrapped and bridged by dense MWCNTs, forming complex sliding conductive networks. Therefore, the strain sensor possesses an impressive sensitivity (gauge factor = 646) and 40% response range. Moreover, a fast response time of 280 ms and detection limit of 0.05% are achieved. The high performance enables good prospects in human detection, like human movement and pulse signals for healthcare. It is also applied to wearable smart data gloves, in which the CNN algorithm is utilized to identify 15 gestures, and the final recognition rate is up to 95%. This comprehensive performance strain sensor is designed for a wide array of human body detection applications and wearable intelligent systems.

Real-time target recognition with all-optical neural networks for ghost imaging.

The generation and structural characteristics of random speckle patterns impact the implementation and imaging quality of computational ghost imaging. Their modulation is limited by traditional electronic hardware. We aim to address this limitation using the features of an all-optical neural network. This work proposes a real-time target recognition system based on an all-optical diffraction deep neural network for ghost imaging. We use a trained neural network to perform pure phase modulation on visible light, and directly complete the target recognition task by detecting the maximum value of light intensity signals at different positions. We optimized the system by simulating the effects of parameters, such as the number of layers of the network, photosensitive pixel, unit area etc., on the final recognition performance, and the accuracy of target recognition reached 91.73%. The trained neural network is materialised by 3D printing technology and experiments confirmed that the system successfully performs real-time target recognition at a low sampling rate of 1.25%. It also verified the feasibility and noise resistance of the system in practical application scenarios.

Suppressing the Maintenance of Information in Working Memory Alters Long-term Memory Traces.

The sensory recruitment hypothesis conceptualizes information in working memory as being activated representations of information in long-term memory. Accordingly, changes made to an item in working memory would be expected to influence its subsequent retention. Here, we tested the hypothesis that suppressing information from working memory, which can reduce short-term access to that information, may also alter its long-term neural representation. We obtained fMRI data (n = 25; 13 female / 12 male participants) while participants completed a working memory removal task with scene images as stimuli, followed by a final surprise recognition test of the examined items. We applied a multivariate pattern analysis to the data to quantify the engagement of suppression on each trial, to track the contents of working memory during suppression, and to assess representational changes afterward. Our analysis confirms previous reports that suppression of information in working memory involves focused attention to target and remove unwanted information. Furthermore, our findings provide new evidence that even a single dose of suppression of an item in working memory can (if engaged with sufficient strength) produce lasting changes in its neural representation, particularly weakening the unique, item-specific features, which leads to forgetting. Our study sheds light on the underlying mechanisms that contribute to the suppression of unwanted thoughts and highlights the dynamic interplay between working memory and long-term memory.

GRNet: a graph reasoning network for enhanced multi-modal learning in scene text recognition

Abstract Recent advancements in scene text recognition have predominantly focused on leveraging textual semantics. However, an over-reliance on linguistic priors can impede a model’s ability to handle irregular text scenes, including non-standard word usage, occlusions, severe distortions, or stretching. The key challenges lie in effectively localizing occlusions, perceiving multi-scale text, and inferring text based on scene context. To address these challenges and enhance visual capabilities, we introduce the Graph Reasoning Model (GRM). The GRM employs a novel feature fusion method to align spatial context information across different scales, beginning with a feature aggregation stage that extracts rich spatial contextual information from various feature maps. Visual reasoning representations are then obtained through graph convolution. We integrate the GRM module with a language model to form a two-stream architecture called GRNet. This architecture combines pure visual predictions with joint visual-linguistic predictions to produce the final recognition results. Additionally, we propose a dynamic iteration refinement for the language model to prevent over-correction of prediction results, ensuring a balanced contribution from both visual and linguistic cues. Extensive experiments demonstrate that GRNet achieves state-of-the-art average recognition accuracy across six mainstream benchmarks. These results highlight the efficacy of our multi-modal approach in scene text recognition, particularly in challenging scenarios where visual reasoning plays a crucial role.

Transformer-based end-to-end attack on text CAPTCHAs with triplet deep attention

Websites frequently use text-based captcha images to distinguish whether the user is a person or not. Previous research mainly focuses on different training strategies and neglects the characteristics of the text-based captcha images themselves, resulting in low accuracy. For text-based captcha images characterized by rotation, distortion, and non-character elements, we propose an end-to-end attack using a Transformer-based method with triplet deep attention. Firstly, the features of text-based captchas are extracted using ResNet45 with triplet deep attention module and Transformer encoder. The TDA module is capable of learning rotational and distortion features of characters. Subsequently, based on self-attention mechanism, design query, key, and value, and adopt the query enhancement module to enhance the query. The query enhancement module can strengthen character localization and reduce attention drift towards non-character elements. Finally, the feature maps are transformed into probabilities of character for the final text recognition. Experiments are conducted on captcha datasets based on Roman characters from 9 popular websites, achieving average word accuracy of 91.14%. To evaluate the performance of our method on data with small samples, experiments are conducted different scales of training data. Additionally, we use the method on Chinese text-based captcha tasks and achieve average word accuracy of 99.60%. The effectiveness of the method is also explored under conditions of lack of illumination and scene text recognition, where background interference is present.

False memories in bilinguals: Integration of information across languages and limits on proficiency effects.

Research on false memory in bilinguals using the DRM task has shown that false memories transfer across languages, but comparisons to within-language conditions have yielded mixed results. In two experiments, Spanish-English bilinguals completed standardized language assessments and a DRM task. Experiment 1 (N = 96) had several study-recall cycles before a final recognition test, and Experiment 2 (N = 72) only tested recognition. Relative to within-language conditions, more critical lures were recalled when the language changed from study to test and when words were studied in mixed-language sequences. With no prior recall test, the rate of critical lure recognition did not differ across language conditions. Language proficiency was not associated with the false-memory effects. Associations of false and veridical memory were negative in recall and positive in recognition. Overall, the findings indicate that proficient bilinguals can integrate information across their languages via a shared semantic network to form false memories.

ChemReco: automated recognition of hand-drawn carbon–hydrogen–oxygen structures using deep learning

Chemical molecular structures are a direct and convenient means of expressing chemical knowledge, playing a vital role in academic communication. In chemistry, hand drawing is a common task for students and researchers. If we can convert hand-drawn chemical molecular structures into machine-readable formats, like SMILES encoding, computers can efficiently process and analyze these structures, significantly enhancing the efficiency of chemical research. Furthermore, with the progress of educational technology, automated grading is gaining popularity. When machines automatically recognize chemical molecular structures and assess the correctness of the drawings, it offers great convenience to teachers. We created ChemReco, a tool designed to identify chemical molecular structures involving three atoms: C, H, and O, providing convenience for chemical researchers. Currently, there are limited studies on hand-drawn chemical molecular structures. Therefore, the primary focus of this paper is constructing datasets. We propose a synthetic image method to rapidly generate images resembling hand-drawn chemical molecular structures, enhancing dataset acquisition efficiency. Regarding model selection, the hand-drawn chemical molecule structural recognition model developed in this article achieves a final recognition accuracy of 96.90%. This model employs the encoder-decoder architecture of EfficientNet + Transformer, demonstrating superior performance compared to other encoder-decoder combinations.

Multi-label recognition in open driving scenarios based on bipartite-driven superimposed dynamic graph

The multi-label image recognition task is widely prevalent in real-world scenarios. Overcoming the issue of overlapping and densely packed objects in complex scenes is crucial. For instance, in traffic scenarios, there are overlaps and close proximity among pedestrians, various types of vehicles, and signage. However, a primary obstacle in leveraging label relationships to enhance image classification lies in effectively integrating label semantic topology information with the image data itself. In this paper, we propose a novel framework, the Bipartite-driven Superimposed Dynamic Graph Convolutional Network (Bi-SDNet), augmented with Mapping Alignment Module (MAM) and Semantic Decoupling Module(SDM). Our approach initially decomposes input features into representations capable of discerning category label semantics at multiple scales, facilitated by MAM and SDM modules. Furthermore, through the meticulously designed Superimposed Dynamic Graph, we adeptly capture content-aware category relationships for each image, effectively modeling the relationships between these representations for the final recognition task. We conducted extensive experiments on publicly available benchmark datasets and the traffic scene dataset WZ-traffic. The model achieved an impressive 87.5% mean average precision (mAP) on the MS-COCO dataset and a commendable 91% mAP on the WZ-traffic dataset. Our research introduces novel techniques and significant breakthroughs in this field, furnishing powerful tools for enhancing model performance.

Intelligent Monitoring System for Highway Roadbed Based on Combination Neural Network Algorithm

To solve the problem of the frequent occurrence of roadbed faults, we studied the highway roadbed intelligent monitoring system based on a combined neural network algorithm. Based on the embedded system, with a variety of sensors, we completed the construction of the roadbed monitoring system. In the selection of the data processing algorithm model, the combined neural network algorithm based on an artificial immune algorithm and probabilistic neural network (PNN) is selected. The accurate acquisition of data characteristics is realized by data preprocessing, data smoothing and data fitting. Through experimental verification, the accuracy of the research model in identifying roadbed settlements has been improved by about 5% compared to traditional models. Furthermore, the processing time of the model has been shortened by about 19.5%, proving the effectiveness of the model. In terms of fault identification, compared with other classic models, the final recognition accuracy of this model reached 96.7%, far exceeding the comparison model. This provides new ideas for the monitoring and protection of roadbed faults.

Pathological Priors Inspired Network for Vertebral Osteophytes Recognition.

Automatic vertebral osteophyte recognition in Digital Radiography is of great importance for the early prediction of degenerative disease but is still a challenge because of the tiny size and high inter-class similarity between normal and osteophyte vertebrae. Meanwhile, common sampling strategies applied in Convolution Neural Network could cause detailed context loss. All of these could lead to an incorrect positioning predicament. In this paper, based on important pathological priors, we define a set of potential lesions of each vertebra and propose a novel Pathological Priors Inspired Network (PPIN) to achieve accurate osteophyte recognition. PPIN comprises a backbone feature extractor integrating with a Wavelet Transform Sampling module for high-frequency detailed context extraction, a detection branch for locating all potential lesions and a classification branch for producing final osteophyte recognition. The Anatomical Map-guided Filter between two branches helps the network focus on the specific anatomical regions via the generated heatmaps of potential lesions in the detection branch to address the incorrect positioning problem. To reduce the inter-class similarity, a Bilateral Augmentation Module based on the graph relationship is proposed to imitate the clinical diagnosis process and to extract discriminative contextual information between adjacent vertebrae in the classification branch. Experiments on the two osteophytes-specific datasets collected from the public VinDr-Spine database show that the proposed PPIN achieves the best recognition performance among multitask frameworks and shows strong generalization. The results on a private dataset demonstrate the potential in clinical application. The Class Activation Maps also show the powerful localization capability of PPIN. The source codes are available in https://github.com/Phalo/PPIN.

Marginal Subspace Learning With Group Low-Rank for Unsupervised Domain Adaptation.

Unsupervised domain adaptation is intended to construct a reliable model for the unlabeled target samples using the well-labeled but differently distributed source samples. To tackle the domain shift issue, learning domain-invariant feature representations across domains is important, and most of the existing methods have concentrated on this goal. However, these methods rarely take into consideration the group discriminability of the feature representation, which is detrimental to the final recognition. Therefore, this article proposes a novel unsupervised domain adaptation method, named marginal subspace learning with group low-rank (MSL-GLR), to extract both domain-invariant and discriminative feature representations. Specifically, MSL-GLR uses the retargeting strategy to relax the regression matrix, such that the regression values would be forced to satisfy a margin maximization criterion for the requirement of correct classification. Moreover, MSL-GLR imposes a class-induced low-rank constraint, which enables the samples of each class to be located in their respective subspace. In this way, the distance between samples from the same class can be decreased and the discriminant ability of the projection is greatly improved. Furthermore, with the help of alternating direction method of multipliers (ADMM), an efficient algorithm is presented to solve the resulting optimization problem. Finally, the effectiveness of the proposed MSL-GLR is demonstrated by comprehensive evaluations on multiple domain adaptation benchmark datasets.

Quantum-inspired vortex search algorithm with deep neural networks for multi-pose facial expression recognition

The rapid expansion of artificial intelligence technologies has enabled machines to comprehend emotional intelligence. Among various indicators, facial expressions serve as an effective medium for understanding emotions. The concept of facial expression recognition (FER) relies heavily on the accurate and robust features available. Initially, the method of three-channel convolutional neural networks (TC-CNN) is adapted to extract facial features. However, only extracting the features is insufficient, the optimization of the extracted features is crucial to determining precise and robust features. This research work focuses on the optimization of the features using the quantum-inspired vortex search algorithm (QVSA). The QVSA integrates the attributes of Q-bits into the vortex search algorithm (VSA), optimizing the features by using the Q-bits to determine the vortex center on the Bloch sphere. The Q-bit attributes also improve the diversity of the features and help to avoid the premature convergence of the VSA. The final recognition of the facial expressions is performed using the deep neural network method of ResNet101v2. The experiments for facial expression recognition are performed on the datasets of RaFD and KDEF, which include different facial positions such as front pose, diagonal pose and profile pose. Performance comparisons demonstrate the effectiveness of the proposed system over state-of-the-art facial expression techniques.

Basketball Sports Posture Recognition Technology Based on Improved Graph Convolutional Neural Network

Basketball has rapidly developed in recent years. Analysis of various moves in basketball can provide technical references for professional players and assist referees in judging games. Traditional technology can no longer provide modern basketball players with theoretical support. Therefore, using intelligent methods to recognize human body postures in basketball was a relatively innovative approach. To be able to recognize the basketball sports posture of players more accurately, the experiment proposes a basketball stance recognition model based on enhanced graph convolutional networks (GCN), that is, the basketball stance recognition model based on enhanced GCN and spatial temporal graph convolutional network (ST-GCN) model. This model combines the respective advantages of the GCN and temporal convolutional network and can handle graph-structured data with time-series relationships. The ST-GCN can be further deduced by realizing the convolution operation of the graph structure and establishing a spatiotemporal graph convolution model for the posture sequence of a person’s body. A dataset of technical basketball actions is constructed to verify the effectiveness of the ST-GCN model. The final experimental findings indicated that the final recognition accuracy of the ST-GCN model for basketball postures was approximately 95.58%, whereas the final recognition accuracy of the long short term memory + multiview re-observation skeleton action recognition (LSTM+MV+AC) model was about 93.65%.

The history of the development of French Savat boxing

This article highlights the history of the development of French Savat boxing. It is indicated that French boxing savat is a spectacular form of martial arts that is gaining popularity in Ukraine, and therefore the question of studying the history of its development is relevant. The purpose of the article is to get acquainted with the history of the development of French Savat boxing. The final recognition of the respectability of sawat came in 1924, when it was included as a demonstration sport at the Olympic Games in Paris. In 2008, Savat was recognized by the International University Sports Federation (FISU) - this recognition allows the World Federation to hold official university world championships; the first was held in Nantes, France in 2010. The twenty-fifth anniversary of the founding of the International Savat Federation in March 2010 was marked by a visit to Lausanne to meet the President of the International Olympic Committee, Jacques Rogge. FISav President Gilles Le Dugot received a souvenir with the image of the Olympic rings. In April 2010, the International Sawat Federation was accepted as a member of SportAccord (formerly known as AGFIS) - a major step forward towards Olympic inclusion. However, the history of the development of French Savat boxing was not conducted by Ukrainian scientists. Studying the specifics of this sport will allow a deeper understanding of not only the sport itself, but also the peculiarities of the worldview and aesthetic preferences of the French people. An in-depth study of the cultural history of Savat through a historical lens will help to more vividly present the uniqueness and value of this tradition. This can interest and encourage young people to engage in French boxing, not only young men, but also girls, and it will inspire young athletes to reach the heights of sportsmanship. Conclusion. If the Europeans discovered the French boxing savat a long time ago and are developing it at a high professional and technical level, this sport is relatively young in Ukraine. Therefore, the scope for research in this area is large. French boxing sawat is an interesting and spectacular sport, which owes its technical elements. Therefore, it is worth paying great attention to the study of various aspects of this type of martial arts.

Discriminative Segment Focus Network for Fine-grained Video Action Recognition

Fine-grained video action recognition aims at identifying minor and discriminative variations among fine categories of actions. While many recent action recognition methods have been proposed to better model spatio-temporal representations, how to model the interactions among discriminative atomic actions to effectively characterize inter-class and intra-class variations has been neglected, which is vital for understanding fine-grained actions. In this work, we devise a Discriminative Segment Focus Network (DSFNet) to mine the discriminability of segment correlations and localize discriminative action-relevant segments for fine-grained video action recognition. Firstly, we propose a hierarchic correlation reasoning (HCR) module which explicitly establishes correlations between different segments at multiple temporal scales and enhances each segment by exploiting the correlations with other segments. Secondly, a discriminative segment focus (DSF) module is devised to localize the most action-relevant segments from the enhanced representations of HCR by enforcing the consistency between the discriminability and the classification confidence of a given segment with a consistency constraint. Finally, these localized segment representations are combined with the global action representation of the whole video for boosting final recognition. Extensive experimental results on two fine-grained action recognition datasets, i.e., FineGym and Diving48, and two action recognition datasets, i.e., Kinetics400 and Something-Something, demonstrate the effectiveness of our approach compared with the state-of-the-art methods.

A Multi-Objective Pigeon-Inspired Optimization Algorithm for Community Detection in Complex Networks

Community structure is a very interesting attribute and feature in complex networks, which has attracted scholars’ attention and research on community detection. Many single-objective optimization algorithms have been migrated and modified to serve community detection problems. Due to the limitation of resolution, the final algorithm implementation effect is not ideal. In this paper, a multi-objective community detection method based on a pigeon-inspired optimization algorithm, MOPIO-Net, is proposed. Firstly, the PIO algorithm is discretized in terms of the solution space representation, position, and velocity-updating strategies to adapt to discrete community detection scenarios. Secondly, by minimizing the two objective functions of community score and community fitness at the same time, the community structure with a tight interior and sparse exterior is obtained. Finally, for the misclassification caused by boundary nodes, a mutation strategy is added to improve the accuracy of the final community recognition. Experiments on synthetic and real networks verify that the proposed algorithm is more accurate in community recognition compared to 11 benchmark algorithms, confirming the effectiveness of the proposed method.

MULTIMODAL SYSTEM FOR FACIAL EMOTION RECOGNITION BASED ON DEEP LEARNING

Emotions are one of the main ways of communication between people and of expressing attitudes towards objects, products, services, etc. They are divided to verbal and non-verbal classes. Human speech and intonation belong to the first class, and to the second (non-verbal) facial and body emotions, known as body language. The subject of this report is the development of multimodal deep learning system intended to recognize facial and body emotions and their relationship with the scene (weather) in which they occur. It is based on three deep learning neural networks (DNN) each one for recognition of facial emotion, body emotion and weather. Combining their results, we improve significantly the final facial emotion recognition (FER) results.