Network For Recognition Research Articles

Manual Acupuncture Manipulation Recognition Method via Interactive Fusion of Spatial Multiscale Motion Features

Manual acupuncture manipulation (MAM) is essential in traditional Chinese medicine treatment. MAM action recognition is important for junior acupuncturists’ training and education; however, there are obvious personalized differences in hand gestures among expert acupuncturists for the same type of MAM. In addition, during the MAM operations, the magnitude and frequency of the expert acupuncturists’ hand shape and relative needle-holding finger position changes are tiny and fast, resulting in difficulties in observing MAM action details. Thus, we propose a Spatial Multiscale Interactive Fusion MAM Recognition Network to solve the difficulties in MAM recognition. First, this paper presents an optical flow-based hand motion contour global feature extraction method for acupuncture hand shape. Second, to explore the motion rule between the needle-holding fingers during the MAM operations, we design a quantitative description method of the relative motion of the needle-holding fingers: an “interactive attention module,” which achieves feature fusion and mines the correlation between different scales of MAM action features. Finally, the proposed MAM recognition method was validated by 20 acupuncturists from the Beijing University of Traditional Chinese Medicine and 10 from the Beijing Zhongguancun Hospital who participated in the MAM video signal collection. The proposed recognition method achieves the highest average validation accuracy of 95.3% and the highest test accuracy of 96.0% for four typical MAMs, proving its feasibility and effectiveness.

Ring artifacts suppression for X-ray CT images by fusion of dual-domain images based on improved UNet

The ring artifact is a typical artifact in X-ray Computed Tomography (CT) images, the existence of which can reduce the image's quality, change the structure and details and affects the interpretation of image information. This paper conducts research from two aspects: tomographic images and sinograms, and designs a dual-domain artifact suppression network. By introducing the Inception module and the residual module, the network's ability to extract features is improved, and the gradient diffusion is suppressed. By customizing the loss function, the network's recognition of ring artifacts is improved and the network's convergence is accelerated. Finally, the outputs of the dual-domain are fused based on the Nonsubsampled Contourlet Transform (NSCT) fusion theory to improve the quality of the image. In addition, three experiments of CT scanning are designed based on microfocus cone-beam CT system to verify the effectiveness. The model is trained and tested using real datasets. Peak Signal-to-Noise Ratio (PSNR), Structural Similarity (SSIM), and Mean Squared Error (MSE) are employed as evaluation indicators of image quality. Finally, the proposed method is compared with typical algorithms for artifact suppression. The experimental results show that the method effectively protects the structure information in CT images while suppressing ring artifacts to the greatest extent. The PSNR, SSIM, and MSE of the processed images can respectively reach 36.167 dB, 0.9847 and 6.87 × e-4.

Intent Recognition and slot filling for ecommerce chatbot

With the continuous evolution of natural language processing (NLP), conversational agents have become pivotal in enhancing user engagement and satisfaction in the e-commerce domain. This paper presents a comprehensive study and implementation of intent recognition and slot filling techniques tailored for an e-commerce chatbot. Leveraging a dataset of 8,000 samples, we employed a Convolutional Neural Network (CNN) combined with Long Short-Term Memory (LSTM) networks for intent recognition, achieving an accuracy of 87%. For slot filling, we utilized the Bidirectional Encoder Representations from Transformers (BERT) model, also attaining an accuracy of 87%. The chatbot seamlessly integrates with an e-commerce database, using OpenAI's GPT-3 to generate natural language responses from query results. Our system demonstrates significant advancements in processing user queries, generating precise database queries, and providing coherent and relevant responses, thereby enhancing the overall user experience. This work aims to equip researchers and practitioners with insights into the methodologies and challenges in developing sophisticated e-commerce chatbots, fostering further innovation in this field. Key Words: Intent Recognition, Slot Filling, Convolutional Neural Network (CNN), Long Short-Term Memory (LSTM), Bidirectional Encoder Representations from Transformers (BERT), GPT-3, Natural Language Understanding (NLU), Chatbot

MCCM: multi-scale feature extraction network for disease classification and recognition of chili leaves.

Currently, foliar diseases of chili have significantly impacted both yield and quality. Despite effective advancements in deep learning techniques for the classification of chili leaf diseases, most existing classification models still face challenges in terms of accuracy and practical application in disease identification. Therefore, in this study, an optimized and enhanced convolutional neural network model named MCCM (MCSAM-ConvNeXt-MSFFM) is proposed by introducing ConvNeXt. The model incorporates a Multi-Scale Feature Fusion Module (MSFFM) aimed at better capturing disease features of various sizes and positions within the images. Moreover, adjustments are made to the positioning, activation functions, and normalization operations of the MSFFM module to further optimize the overall model. Additionally, a proposed Mixed Channel Spatial Attention Mechanism (MCSAM) strengthens the correlation between non-local channels and spatial features, enhancing the model's extraction of fundamental characteristics of chili leaf diseases. During the training process, pre-trained weights are obtained from the Plant Village dataset using transfer learning to accelerate the model's convergence. Regarding model evaluation, the MCCM model is compared with existing CNN models (Vgg16, ResNet34, GoogLeNet, MobileNetV2, ShuffleNet, EfficientNetV2, ConvNeXt), and Swin-Transformer. The results demonstrate that the MCCM model achieves average improvements of 3.38%, 2.62%, 2.48%, and 2.53% in accuracy, precision, recall, and F1 score, respectively. Particularly noteworthy is that compared to the original ConvNeXt model, the MCCM model exhibits significant enhancements across all performance metrics. Furthermore, classification experiments conducted on rice and maize disease datasets showcase the MCCM model's strong generalization performance. Finally, in terms of application, a chili leaf disease classification website is successfully developed using the Flask framework. This website accurately identifies uploaded chili leaf disease images, demonstrating the practical utility of the model.

SDRnet: A Deep Fusion Network for ISAR Ship Target Recognition Based on Feature Separation and Weighted Decision

Existing methods for inverse synthetic aperture radar (ISAR) target recognition typically rely on a single high-resolution radar signal type, such as ISAR images or high-resolution range profiles (HRRPs). However, ISAR images and HRRP data offer representations of targets across different aspects, each containing valuable information crucial for radar target recognition. Moreover, the process of generating ISAR images inherently facilitates the acquisition of HRRP data, ensuring timely data collection. Therefore, to fully leverage the different information from both HRRP data and ISAR images and enhance ISAR ship target recognition performance, we propose a novel deep fusion network named the Separation-Decision Recognition network (SDRnet). First, our approach employs a convolutional neural network (CNN) to extract initial feature vectors from ISAR images and HRRP data. Subsequently, a feature separation module is employed to derive a more robust target representation. Finally, we introduce a weighted decision module to enhance overall predictive performance. We validate our method using simulated and measured data containing ten categories of ship targets. The experimental results confirm the effectiveness of our approach in improving ISAR ship target recognition.

EEG Emotion Recognition Network Based on Attention and Spatiotemporal Convolution.

Human emotions are complex psychological and physiological responses to external stimuli. Correctly identifying and providing feedback on emotions is an important goal in human-computer interaction research. Compared to facial expressions, speech, or other physiological signals, using electroencephalogram (EEG) signals for the task of emotion recognition has advantages in terms of authenticity, objectivity, and high reliability; thus, it is attracting increasing attention from researchers. However, the current methods have significant room for improvement in terms of the combination of information exchange between different brain regions and time-frequency feature extraction. Therefore, this paper proposes an EEG emotion recognition network, namely, self-organized graph pesudo-3D convolution (SOGPCN), based on attention and spatiotemporal convolution. Unlike previous methods that directly construct graph structures for brain channels, the proposed SOGPCN method considers that the spatial relationships between electrodes in each frequency band differ. First, a self-organizing map is constructed for each channel in each frequency band to obtain the 10 most relevant channels to the current channel, and graph convolution is employed to capture the spatial relationships between all channels in the self-organizing map constructed for each channel in each frequency band. Then, pseudo-three-dimensional convolution combined with partial dot product attention is implemented to extract the temporal features of the EEG sequence. Finally, LSTM is employed to learn the contextual information between adjacent time-series data. Subject-dependent and subject-independent experiments are conducted on the SEED dataset to evaluate the performance of the proposed SOGPCN method, which achieves recognition accuracies of 95.26% and 94.22%, respectively, indicating that the proposed method outperforms several baseline methods.

Endangered Bird Species Classification using Machine Learning

The conservation of endangered bird species is a critical aspect of biodiversity preservation. Traditional methods of identifying and classifying these species are often labor-intensive and time-consuming. In recent years, advancements in machine learning have offered promising alternatives for enhancing the accuracy and efficiency of such tasks. This paper explores the application of various machine learning algorithms to the classification of endangered bird species. By leveraging a dataset comprising images and audio recordings of bird calls, we train and evaluate models such as Convolutional Neural Networks (CNNs) for image recognition and Recurrent Neural Networks (RNNs) for audio classification. Our results demonstrate that these models can achieve high accuracy rates, significantly surpassing traditional methods. Furthermore, we discuss the importance of feature selection, data augmentation, and the integration of multimodal data in improving model performance. The findings underscore the potential of machine learning to revolutionize wildlife conservation efforts, providing a scalable and robust tool for the timely identification and protection of endangered bird species.

Fault Detection and Normal Operating Condition in Power Transformers via Pattern Recognition Artificial Neural Network

Aging, degradation, or damage to internal insulation materials often contribute to transformer failures. Furthermore, combustible gases can be produced when these insulation materials experience thermal or electrical stresses. This paper presents an artificial neural network for pattern recognition (PRN) to classify the operating conditions of power transformers (normal, thermal faults, and electrical faults) depending on the combustible gases present in them. Two network configurations were presented, one with five and the other with ten neurons in the hidden layer. The main advantage of applying this model through artificial neural networks is its ability to capture the nonlinear characteristics of the samples under study, thus avoiding the need for iterative procedures. The effectiveness and applicability of the proposed methodology were evaluated on 815 real data samples. Based on the results, the PRN performed well in both training and validation (for samples that were not part of the training), with a mean squared error (MSE) close to expected (0.001). The network was able to classify the samples with a 98% accuracy rate of the 815 samples presented and with 100% accuracy in validation, showing that the methodology developed is capable of acting as a tool for diagnosing the operability of power transformers.

Lightweight and real-time semantic segmentation of UAV traffic videos based on siamese network for keyframe recognition

Lightweight and real-time semantic segmentation of UAV traffic videos based on siamese network for keyframe recognition

Maize leaf disease recognition based on improved MSRCR and OSCRNet

Accurate identification of maize pests and diseases plays a crucial role in the growth and yield of maize. The recognition method of maize leaf diseases based on deep learning is currently a research hotspot. In practical applications, it is difficult to ensure a high recognition accuracy due to the impact of environmental factors (such as light intensity, background noise) and insufficient training data. To address this challenge, this paper proposes a maize image recognition method based on image enhancement and OSCRNet. Firstly, a maize leaf image enhancement framework and algorithm based on improved MSRCR were designed to enhance the characteristics of the maize leaf in a complex environment and to solve the problems of high noise and blur of maize images. Subsequently, the OSCRNet maize leaf recognition network model was established, which uses octave convolution with characteristics to accelerate network training and reduce unnecessary redundant spatial information in maize leaf images. Additionally, a self-calibrated convolution with multi-scale features was employed to realize the interactions of different feature information in the maize leaf images, enhance feature extraction, and solve the problems of similarity of maize disease features and easy learning disorders. Concurrently, batch normalization was employed to prevent network overfitting and enhance the robustness of the model. The experimental results showed that the highest identification accuracy of the proposed method for rust, gray leaf disease, northern wilt disease, and healthy maize are 96.67%, 94.86%, 92.81%, and 95.63%, respectively. Our methods were beneficial in solving the problems of slow efficiency, low accuracy and insufficient training data, and also outperformed other comparison models. The proposed method demonstrates strong robustness for maize disease images collected in the natural environment, providing a reference for the intelligent diagnosis of other plant leaf diseases.

A hardware-implemented neural network for the analysis of handwritten numerals

This study presents a specialized hardware-accelerated neural network tailored for the recognition of handwritten digits in 28x28 pixel grayscale images. Employing the perceptron model, our single-layer neural network is composed of 10 neurons, each handling inputs from all pixels to generate an output. The digit recognition is determined by the neuron with the highest output value. Implemented in synthesizable Verilog, the design complies with a constraint of 350 multipliers. To achieve this, this paper employs a combination of parallel processing and pipelining, breaking down the 785 multiplications needed for each digit into 8 stages, simultaneously processing 98 data points per clock cycle. In testbench evaluations, the final design exhibits impressive performance, successfully recognizing the majority of the provided images, and attaining a remarkable 99% accuracy rate, all with a minimal delay of just 115 clocks. This accomplishment is achieved using only 99 multipliers and 107 adders, showcasing the efficiency and effectiveness of our hardware-accelerated neural network for handwritten digit recognition.

DSTC-Net: differential spatio-temporal correlation network for similar action recognition

DSTC-Net: differential spatio-temporal correlation network for similar action recognition

AM-MSFF: A Pest Recognition Network Based on Attention Mechanism and Multi-Scale Feature Fusion.

Traditional methods for pest recognition have certain limitations in addressing the challenges posed by diverse pest species, varying sizes, diverse morphologies, and complex field backgrounds, resulting in a lower recognition accuracy. To overcome these limitations, this paper proposes a novel pest recognition method based on attention mechanism and multi-scale feature fusion (AM-MSFF). By combining the advantages of attention mechanism and multi-scale feature fusion, this method significantly improves the accuracy of pest recognition. Firstly, we introduce the relation-aware global attention (RGA) module to adaptively adjust the feature weights of each position, thereby focusing more on the regions relevant to pests and reducing the background interference. Then, we propose the multi-scale feature fusion (MSFF) module to fuse feature maps from different scales, which better captures the subtle differences and the overall shape features in pest images. Moreover, we introduce generalized-mean pooling (GeMP) to more accurately extract feature information from pest images and better distinguish different pest categories. In terms of the loss function, this study proposes an improved focal loss (FL), known as balanced focal loss (BFL), as a replacement for cross-entropy loss. This improvement aims to address the common issue of class imbalance in pest datasets, thereby enhancing the recognition accuracy of pest identification models. To evaluate the performance of the AM-MSFF model, we conduct experiments on two publicly available pest datasets (IP102 and D0). Extensive experiments demonstrate that our proposed AM-MSFF outperforms most state-of-the-art methods. On the IP102 dataset, the accuracy reaches 72.64%, while on the D0 dataset, it reaches 99.05%.

JGULF: Joint global and unilateral local feature network for micro-expression recognition

Micro-expression is a subtle facial movement that is fleeting and manifest in localized areas, making it difficult for the human eye to detect and recognize it. Although algorithms that extract facial features from specific regions or the entire face have shown potential, the classification of micro-expressions using features from symmetrical left and right regions can be challenging in the presence of unilateral movements. This can ultimately affect the performance of micro-expression recognition. To address this issue, we propose a network called Joint Global and Unilateral Local Features (JGULF) for micro-expression recognition. Initially, we employ a Convolutional Neural Network (CNN) and an adjusted Vision Transformer (ViT) model to extract global features from micro-expressions. The local feature extraction module is designed based on global features. The facial features are divided into multiple local regions with varying scales. After that, local feature learning and selection are performed to filter out unilateral local features related to micro-expression movements efficiently. Finally, global and local features are combined to classify micro-expressions. Through comprehensive experimental validation, our algorithm achieves state-of-the-art classification performance on the SMIC, CASMEII, and SAMM micro-expression datasets, demonstrating the effectiveness of combining global features and selecting local features.

GMSN: An efficient multi-scale feature extraction network for gait recognition

In current gait recognition methods, researchers predominantly focus on gait information at specific spatial scales, with a tendency to overlook information variances across different scales. Additionally, from observation, we found variations in the spatiotemporal information offered by different human body parts. To address these issues, we present a Multi-scale Network for Gait Recognition (GMSN), which aims to highlight key body parts in a more discriminative gait representation. GMSN consists of two key modules: the Multi-scale Feature Extractor (MSFE) and the Part-based Horizontal Mapping (PHM). MSFE employs multi-scale parallel convolutional networks to comprehensively learn features across different scales, capturing both local details and global information for an enriched representation of gait. Meanwhile, PHM focuses on enhancing the learning of crucial body parts that provide clear contours and movement patterns. Experiments on three public datasets demonstrate that our approach attains state-of-the-art recognition accuracy. On the CASIA-B dataset, our model achieves rank-1 accuracies of 98.2 %, 96.0 %, and 87.0 % under normal walking, bag-carrying, and coat-wearing conditions, respectively. On the OU-MVLP and GREW datasets, it achieves a rank-1 accuracy of 90.4 % and 50.6 %, respectively. Also, it can achieve relatively stable results when adding square occlusions to the test samples.

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.

FER-BHARAT: a lightweight deep learning network for efficient unimodal facial emotion recognition in Indian context

Humans' ability to manage their emotions has a big impact on their ability to plan and make decisions. In order to better understand people and improve human–machine interaction, researchers in affective computing and artificial intelligence are investigating the detection and recognition of emotions. However, different cultures have distinct ways of expressing emotions, and the existing emotion recognition datasets and models may not effectively capture the nuances of the Indian population. To address this gap, this study proposes custom-built lightweight Convolutional Neural Network (CNN) models that are optimized for accuracy and computational efficiency. These models are trained and evaluated on two Indian emotion datasets: The Indian Spontaneous Expression Dataset (ISED) and the Indian Semi Acted Facial Expression Database (iSAFE). The proposed CNN model with manual feature extraction provides remarkable accuracy improvement of 11.14% for ISED and 4.72% for iSAFE datasets as compared to baseline, while reducing the training time. The proposed model also surpasses the accuracy produced by pre-trained ResNet-50 model by 0.27% ISED and by 0.24% for the iSAFE dataset with significant improvement in training time of approximately 320 s for ISED and 60 s for iSAFE dataset. The suggested lightweight CNN model with manual feature extraction offers the advantage of being computationally efficient and more accurate compared to pre-trained model making it a more practical and efficient solution for emotion recognition among Indians.

Fragrant: frequency-auxiliary guided relational attention network for low-light action recognition

Fragrant: frequency-auxiliary guided relational attention network for low-light action recognition

PCINet: a Prototype- and Concept-based Interpretable Network for Mutli-scene Recognition

Abstract. With the development of remote sensing techniques, a large number of high-resolution aerial images is now available and benefit many applications. Multi-scene recognition plays a key role in applying remote sensing images to these applications, which refers to predicting multiple scenes coexisted in an aerial image and has attracted an increasing attention. Recently, most researchers tend to invent deep learning-based recognition models and has gained great achievements. However, few efforts have been deployed to explaining the success of deep neural networks in multi-scene recognition. To address this, we introduce concept bottleneck model (CBM) to interpreting model performance and propose a novel network, namely Prototype- and Concept-based Interpretable Network (PCINet), that projects aerial imagery into a prototype-concept memory bank and encode their correlations for explaining how a network can identify coexisting scenes in an aerial image. Specifically, the proposed network mainly consists of two branches: prototype matching that measures similarity scores between image features and scene prototypes, and concept bottleneck branches that aligned image features to textual embeddings and compute their relations with concept embeddings. Afterwards, Outputs are integrated for inferring scene categories. Experimental results show that the model enhances interpretability, providing valuable insights for urban planning and resource management, thereby bridging the gap between deep learning models and practical applications.

Visual smoke recognition based on an inverse-radiating attention pyramid network

In thermal power generation systems, the exhaust gases produced during the process of power generation should be fully burnt before being discharged into the air through a high-altitude chimney, to help maintain safe production and to avoid pollution. A thermal power generation system will emit white smoke when operating normally; otherwise, under abnormal conditions: (1) large amounts of dark smoke may be released into the air due to incomplete combustion of the exhaust gases, leading to hazards to public health and the environment; (2) colorless smoke can be emitted if toxic exhaust gases are directly released into the air, also causing dangerous pollution. To address these issues, we have developed an inverse-radiating attention pyramid network (IRAP-Net) for visual smoke recognition, directed towards monitoring thermal power generation systems. The proposed IRAP-Net exploits various visual smoke features. IRAP-Net is based on three pyramid blocks, composed of 3, 4 and 5 convolution modules, respectively. An attention mechanism is introduced into each pyramid block selective of features. Lastly, an inverse-radiating connection converges all the outputs of the pyramid blocks in a feedforward manner, thereby merging low-level and high-level features. Experiments on a large smoke image database show that the proposed IRAP-Net achieves a recognition rate exceeding 95.5%, superior to other state-of-the-art (SOTA) models.