Museum Relic Image Detection and Recognition Based on Deep Learning

Traditional image recognition algorithms often face problems such as low recognition accuracy and insufficient robustness when facing complex scenes and multi-class image data. To this end, a hierarchical discriminant analysis (HDA)-based image recognition algorithm was proposed, which effectively improves image recognition performance by constructing a multi-scale feature extraction module, principal component analysis (PCA) dimensionality reduction, attention mechanism, and dynamic hierarchical adjustment strategy combined with a hierarchical feature extraction and discrimination model. The experiment was conducted on three public datasets: CIFAR-10, ImageNet subset (selecting 100 categories with a total of 150000 images, based on covering common object categories and moderate data volume for fair validation of algorithm performance), and MNIST. The performance was compared with models such as VGG16, ResNet50, SVM, KNN, Hierarchical CNN, EfficientNet, GoogLeNet, etc. The results indicated that the proposed method had higher recognition accuracy than other comparative algorithms on different datasets, with accuracies exceeding 90%. The proposed method performed better in terms of mean absolute error and root mean squared error. The F1 value curve of the proposed method was located at the top of the coordinate axis, reaching a maximum value of 92.39%, which was 14.56% higher than the lowest value of 78.24% in the EfficientNet model. This algorithm has better recognition accuracy than traditional algorithms on multiple public datasets, and has strong anti-interference ability and robustness, which can provide reference for optimizing the accuracy of image recognition.

With the diversification of network traffic and the popularization of encrypted communication, traditional traffic classification methods have been difficult to cope with the challenges of new encryption protocols and hidden traffic. Deep learning, as a powerful feature extraction and classification tool, has been widely used in network traffic classification. In this paper, a deep network traffic classification model based on multi-scale feature fusion and attention mechanism (MSF-AttNet) is proposed. Features of different time scales are extracted through multi-scale convolution, and the adaptive weighting of channel attention and temporal self-attention mechanisms is combined to significantly improve the classification accuracy and interpretability of the model. Experiments show that MSF-AttNet outperforms traditional methods and other deep learning models on the ISCX VPN-nonVPN dataset.

Both visible and infrared images are important sources of intelligence information on the battlefield, and air‐to‐ground reconnaissance by UAV is an important means to obtain intelligence. However, there are great challenges in ground target detection and tracking, especially in complex battlefield environments. Aiming at the problem of insufficient accuracy of target detection by a single type of sensor in the battlefield environment at this stage, a target detection method by fusion of visible and infrared images is proposed in this paper, which is called ReconnaissanceFusion‐YOLO (RF‐YOLO), and with the help of infrared imagery, it can effectively improve the accuracy of target detection in the case of insufficient light. The performance of target detection in the battlefield is significantly improved by introducing two key innovative modules: dual feature fusion (DFF) module and feature fusion corrector (FFC) module. The DFF module enhances multi‐channel feature fusion through a novel concatenation and channel‐wise attention mechanism, while the FFC module performs feature correction between parallel streams using spatial and channel‐wise weights, addressing noise and uncertainty in different modalities. These modules are integrated on top of a dual‐stream YOLO architecture, allowing for effective fusion of visible and infrared information. RF‐YOLO was trained and evaluated using the FLIR data set, containing 5142 pairs of strictly aligned visible and infrared images. Results demonstrate that RF‐YOLO significantly outperforms benchmark networks in terms of robustness requirements. Specifically, the large model of RF‐YOLO achieves an mAP of 0.831, which is a significant improvement compared to the YOLOv5l inf benchmark's 0.739. This represents an improvement of over 12% in detection accuracy. Additionally, RF‐YOLO offers a Nano version that balances accuracy and speed. The Nano version achieves an mAP of 0.765, while maintaining a model size of only 11.5 MB, making it suitable for deployment on UAV edge computing devices with limited resources. To validate the practical applicability of our approach, this paper successfully implements target detection and tracking on a real UAV's edge computing device using the ROS system and SiameseRPN, combined with the proposed RF‐YOLO. Real‐world flight tests were conducted on an internal playground, demonstrating the effectiveness of our method in actual UAV applications. The system achieved a processing rate of approximately 10 fps at 640 × 640 resolution on an NVIDIA TX2 edge computing device, showcasing its real‐time performance capability in practical scenarios. This study contributes to enhancing UAV‐based battlefield reconnaissance capabilities by improving the accuracy and robustness of target detection and tracking in complex environments. The proposed RF‐YOLO method, along with its successful implementation on a real UAV platform, provides a promising solution for advanced military intelligence gathering and decision‐making support.

Recognizing materials in cultural relic images using computer vision and attention mechanism

In medical imaging diagnosis, accurate segmentation of the knee joint can help doctors better observe and diagnose lesions, thereby improving diagnostic accuracy and treatment effectiveness. Vision Mamba mainly relies on the State Space Model (SSM) for feature modeling, which excels at capturing global contextual information but cannot capture local texture features. Moreover, features of different scales are not effectively integrated, resulting in the model's weak segmentation ability on small-scale tissues (such as cartilage areas). To this end, this study proposed a novel multi-scale Vision Mamba Unet (VM-Unet) framework named MSPF-VM-Unet to perform the segmentation on the femur, tibia, femoral cartilage, and tibial cartilage in knee MRI images. The proposed MSPF-VM-Unet extends VM-Unet by introducing a designed multi-scale pyramid feature extraction network named MPSK, which synergizes multi-resolution feature extraction with channel-space attention. MPSK network enhances multi-scale local feature extraction through Selective Kernel (SK) convolution and pyramid pooling. The network merges the overall context information extracted by the Vision Mamba encoder to achieve the coordinated optimization of a multi-scale hierarchical feature fusion mechanism and global long-range dependency modeling. The results of the comparative experiments on the OAI-ZIB dataset indicate that MSPF-VM-Unet significantly improves the boundary accuracy and regional consistency of the MRI tibiofemoral joint tissue structure.

Online knowledge distillation conducts knowledge transfer among all student models to alleviate the reliance on pre-trained models. However, existing online methods rely heavily on the prediction distributions and neglect the further exploration of the representational knowledge. In this paper, we propose a novel Multi-scale Feature Extraction and Fusion method (MFEF) for online knowledge distillation, which comprises three key components: Multi-scale Feature Extraction, Dual-attention and Feature Fusion, towards generating more informative feature maps for distillation. The multiscale feature extraction exploiting divide-and-concatenate in channel dimension is proposed to improve the multi-scale representation ability of feature maps. To obtain more accurate information, we design a dual-attention to strengthen the important channel and spatial regions adaptively. Moreover, we aggregate and fuse the former processed feature maps via feature fusion to assist the training of student models. Extensive experiments on CIF AR-10, CIF AR-100, and CINIC-10 show that MFEF transfers more beneficial representational knowledge for distillation and outperforms alternative methods among various network architectures

Image recognition and understanding is considered as a remarkable subfield of Artificial Intelligence (AI). In practice, retinal image data have high dimensionality leading to enormous size data. As the morphological retinal image datasets can be analyzed in an expansive and non-invasive way, AI more precisely Deep Learning (DL) methods are facilitating in developing intelligent retinal image analysis tools. The most recently developed DL technique, Convolutional Neural Network (CNN) showed remarkable efficiency in identifying, localizing, and quantifying the complex and hierarchical image features that are responsible for severe cardiovascular diseases. Different deep layered CNN architectures such as LeeNet, AlexNet, and ResNet have been developed exploiting CNN morphology. This wide variety of CNN structures can iteratively learn complex data structures of different datasets through supervised or unsupervised learning and perform exquisite analysis for feature recognition independently to diagnose threatening cardiovascular diseases. In modern ophthalmic practice, DL based automated methods are being used in retinopathy screening, grading, identifying, and quantifying the pathological features to employ further therapeutic approaches and offering a wide potentiality to get rid of ophthalmic system complexity. In this review, the recent advances of DL technologies in retinal image segmentation and feature extraction are extensively discussed. To accomplish this study the pertinent materials were extracted from different publicly available databases and online sources deploying the relevant keywords that includes retinal imaging, artificial intelligence, deep learning and retinal database. For the associated publications the reference lists of selected articles were further investigated.

With the rapid development of deep learning, no-reference image quality assessment (NR-IQA) based on convolutional neural network (CNN) plays an important role in image processing. Currently, most CNN-based NR-IQA methods focus primarily on the global features of images while ignoring detail-rich local features and channel dependencies. In fact, there are subtle differences in detail between distorted and reference images, as well as differences in the contribution of different channels to IQA. Furthermore, multi-scale feature extraction can be used to fuse the detailed information from images with different resolutions, and the combination of global and local features is critical in extracting image features. As a result, in this paper, a multi-scale residual CNN with an attention mechanism (MsRCANet) is proposed for NR-IQA. Specifically, a multi-scale residual block is first used to extract features from distorted images. Then, the residual learning with active weighted mapping strategy and channel attention mechanism is used to further process image features to obtain more abundant information. Finally, the fusion strategy and full connection layer are used to evaluate image quality. The experimental results on four synthetic databases and three in-the-wild IQA databases, as well as cross-database validation results, show that the proposed method has good generalization ability and can be compared with the most advanced methods.

Image target detection and recognition is an important research area in image understanding and computer vision, and has attracted extensive attention of scholars for its wide application. In recent years, with the enlightenment of human vision system, deep learning technology has been put forward, and has become a hot spot. The deep learning model has a multi-layer network structure, which can overcome the shortcomings of the traditional image target recognition methods. In the detection and recognition of image targets, deep learning methods are more and more widely used. This is because the deep learning multilayer network structure can express complex functions in a more concise way than the shallow network, and can learn the deep feature representation. Therefore, this paper uses the method based on deep learning, researches on image target detection and recognition, mainly introduces the region proposal based on the depth of learning based target recognition method of support vector machine algorithm and target detection depth. In this paper, deep learning is applied to image detection and recognition, further to improve the accuracy of image detection and recognition. It has important guiding significance in deep learning and image research.

Single image deraining using multi-scales context information and attention network

Gait recognition is a distinctive biometric technique that can identify pedestrians by their walking patterns from considerable distances. A critical challenge in gait recognition lies in effectively acquiring discriminative spatial-temporal representations from silhouettes that exhibit invariance to disturbances. In this paper, we present a novel gait recognition network by aggregating features in the spatial-temporal and view domains, which consists of two-path spatial-temporal feature fusion module and view embedding module. Specifically, two-path spatial-temporal feature fusion module firstly utilizes multi-scale feature extraction (MSFE) to enrich the input features with multiple convolution kernels of various sizes. Then, frame-level spatial feature extraction (FLSFE) and multi-scale temporal feature extraction (MSTFE) are parallelly constructed to capture spatial and temporal gait features of different granularities and these features are fused together to obtain muti-scale spatial-temporal features. FLSFE is designed to extract both global and local gait features by employing a specially designed residual operation. Simultaneously, MSTFE is applied to adaptively interact multi-scale temporal features and produce suitable motion representations in temporal domain. Taking into account the view information, we introduce a view embedding module to reduce the impact of differing viewpoints. Through the extensive experimentation over CASIA-B and OU-MVLP datasets, the proposed method has achieved superior performance to the other state-of-the-art gait recognition approaches.

Bearing fault diagnosis is vital to safeguard the heath of rotating machinery. It can help to avoid economic losses and safe accidents in time. Effective feature extraction is the premise of diagnosing bearing faults. However, effective features characterizing the health status of bearings are difficult to extract from the raw bearing vibration signals. Furthermore, inefficient feature extraction results in substantial time wastage, making it hard to apply in real time monitoring. A novel feature extraction method for diagnosing bearing faults using multi-scale improved envelope spectrum entropy (MIESE) is proposed in this work. First, bearing vibration signals are analyzed across multiple scales, and improved envelope spectrum entropy (IESE) is extracted from these signals at each scale to form an original feature set. Subsequently, joint approximate diagonalization eigen (JADE) is applied to fuse above feature set for effectively eliminating redundancy and generated a refined feature set. Finally, the newly generated feature set is input into support vector machines (SVM) to effectively diagnose bearing health status. Two cases studies are employed to demonstrate the reliability of the proposed method. The results illustrate the proposed method can improve the stability of extracted features and increase the computational efficiency. Conflict of Interest Statement The authors declare no conflicts of interest.

Recently, deep learning convolutional neural network has been widely used in food recognition and other fields, but there are still some problems such as poor semantic feature extraction ability and low recognition accuracy of Chinese food images. Based on the above reasons, this paper proposed a DenseNet model based on attention mechanism for image recognition of Chinese food. Using the idea of transfer learning, this paper applied the DenseNet pre-training model with excellent classification performance to the Chinese food image dataset. In order to improve the ability of extracting distinguishable features and learning fine-grained features of Chinese food images, this paper proposed an attention mechanism based on the DenseNet-169 model. This paper used the attention module to extract the key semantic feature map of the image, and then carried out the adaptive multi-fine-grained region tailoring. Finally, based on the region weight fusion scheme, the multi-fine-grained region feature map was integrated into an image feature description and sent to DenseNet-169 for recognition, so as to realize the end-to-end image recognition network. Experiments were carried out on VIREO Food-172 dataset, and the results shows that our method achieves considerable and comparable recognition performance to the state-of-art.

The differentiation of benign and malignant skin lesion types is an essential problem in medical images, which facilitates diagnosis and treatment of skin cancer. Traditional deep learning approaches usually depend on the use of the pre-trained models, which are not customized to the specific dataset or task. In this work, to tackle the challenge, a Custom Residual-based Depth-wise Separable Lightweight Inception Model with Custom Attention Mechanism is proposed together with Explainable AI (XAI) techniques. It is written from scratch, using only bottleneck convolutions (1x1) and spatially separable convolutions (1x3,3x1), which makes it really lightweight and efficient. The approach includes developing a new model architecture based on residuals and multi-scale (which known as Inception-style) feature extraction using a novel attention mechanism that guides the network on the important image features. Furthermore, XAI methods such as Grad-CAM, LIME, and SHAP are employed to analyze and explain the model's output, facilitating transparency and reliability in the medical domain. This coursework train and test the model on an accessible skin lesion dataset, using extensive data preprocessing, data augmentation, and class balancing. The results show that the proposed model is superior, with highest accuracy, precision, recall, and F1-score as well as good ROC-AUC and PR-AUC. The XAI visualizations also further confirm the model’s attention on clinically relevant regions of the images, which makes the model more interpretable. In conclusion, a deep learning model that is lightweight, interpretable, and efficient for skin lesion classification is proposed. The combination of PI layers and XAI methodologies guarantees a good level of performance and transparency, thus proving to be an effective instrument for the analysis of medical images.

Land-cover classification based on remote sensing technology has been adopted for decision-making concerning agricultural development, urban planning, and ecosystem protection in arid oasis zones. The semantic segmentation method based on deep learning, as a new paradigm, can effectively overcome the limitations of traditional pixel-based and object-based methods and obtain good classification results from high spatial resolution (HSR) remote sensing images. However, how to extract the exact category boundary and realize the high precision mapping is still a problem. This paper proposes a novel high-resolution feature pyramid attention network (HRFPANet) for land-cover classification. It effectively integrates the advantages of multi-scale feature extraction, attention mechanism, and feature fusion and alleviates boundary inconsistency, roughness, and category fragmentation associated with previous semantic segmentation models. The experimental results show that the mIoU score of HRFPANet is 79.5%, which is 11.5% and 2.6% higher than that of PSPNet and UPerNet, respectively. It proves the proposed model can be used for qualified land-cover mapping in arid oasis zones. Our source code is available at https://github.com/HPU-CPD/HRFPANet.git.