It is an essential measure for workers to wear safety helmets when entering the construction site to prevent head injuries caused by object collision and falling. This paper proposes a lightweight algorithm for helmet-wearing detection based on YOLOV5, which is faster and more robust for helmet detection in natural construction scenarios. In this paper, the MCA attention mechanism is embedded in the backbone network to help the network extract more productive information, reduce the missed detection rate of small helmet objects and improve detection accuracy. In order to ensure the safety of workers in construction, it is necessary to detect whether the construction workers are wearing safety helmets in real-time to achieve monitoring on-site. A channel pruning strategy is proposed on the MCA-YOLOv5 algorithm to compress it, realizing the optimal large-scale model into ultrasmall models for real-time detection on embedded or mobile devices. The experimental results on the public data set show that the model parameter volume is reduced by 87.2%, and the detection speed is increased by 53.5%, even though the MCA-YOLOv5-light reduces the mAP slightly.

It is highly significant from a research standpoint and a valuable practice to identify diseases, symptoms, drugs, examinations, and other medical entities in medical text data to support knowledge maps, question and answer systems, and other downstream tasks that can provide the public with knowledgeable answers. However, when contrasted with other languages like English, Chinese words lack a distinct dividing line, and medical entities have problems such as long length and multiple entity types nesting. Therefore, to address these issues, this study suggests a medical named entity recognition (NER) approach that combines part-of-speech and stroke features. First, the text is fed into the BERT pre-training model to get the semantic representation of the text, while the part-of-speech feature vector is obtained using the part-of-speech dictionary, and the stroke feature of the text is extracted through a convolution neural network (CNN). The word vector is then joined with the part-of-speech and stroke feature vectors, respectively, and input into the BiLSTM and CRF layer for training. Additionally, to balance the disparity in data volume across several types of entities, the class-weighted loss function is included in the loss function. According to the experimental findings, our model’s F1 score on the CCKS2019 dataset reaches 78.65%, and the recognition performance exceeds many existing algorithms.

Deep convolutional neural networks have demonstrated significant performance improvements in face super-resolution tasks. However, many deep learning-based approaches tend to overlook the inherent structural information and feature correlation across different scales in face images, making the accurate recovery of face structure in low-resolution cases challenging. To address this, this paper proposes a method that fuses multi-scale features while preserving the facial structure. It introduces a novel multi-scale residual block (MSRB) to reconstruct key facial parts and structures from spatial and channel dimensions, and utilizes pyramid attention (PA) to exploit non-local self-similarity, improving the details of the reconstructed face. Feature Enhancement Modules (FEM) are employed in the upscale stage to refine and enhance current features using multi-scale features from previous stages. The experimental results on CelebA, Helen and LFW datasets provide evidence that our method achieves superior quantitative metrics compared to the baseline, the Peak Signal-to-Noise Ratio (PSNR) outperforms the baseline by 0.282 dB, 0.343 dB, and 0.336 dB. Furthermore, our method demonstrates improved visual performance on two additional no-reference datasets, Widerface and Webface.

In this paper, a deep learning-based prediction model VMD-GLRT is proposed to address the accuracy problem of service load prediction. The VMD-GLRT model combines Variational Mode Decomposition (VMD) and GRU-LSTM. At the same time, the model incorporates residual networks and self-attentive mechanisms to improve accuracy of the model. The VMD part decomposes the original time series into several intrinsic mode functions (IMFs) and a residual part. The other part uses a GRU-LSTM structure with ResNets and Self-Attention to learn the features of the IMF and the residual part. The model-building process focuses on three main aspects: Firstly, a mathematical model is constructed based on the data characteristics of the service workload. At the same time, VMD is used to decompose the input time series into multiple components to improve the efficiency of the model in extracting features from the data. Secondly, a long and short-term memory (LSTM) network unit is incorporated into the residual network, allowing the network to correct the predictions more accurately and improve the performance of the model. Finally, a self-focus mechanism is incorporated into the model, allowing the model to better capture features over long distances. This improves the dependence of the output vector on these features. To validate the performance of the model, experiences were conducted using open-source datasets. The experimental results were compared with other deep learning and statistical models, and it was found that the model proposed in this paper achieved improvements in mean absolute percentage error (MAPE).

Student behavior recognition in classroom environments is important for teaching quality assessment and intelligent education, yet it remains challenging due to dense student distributions, frequent occlusion, substantial scale variation, and the subtle nature of common classroom activities. To address these issues, this paper proposes RepYOLOv5-SF3D, a cascaded visual perception framework for fine-grained student behavior recognition in complex classroom scenes. The framework integrates a lightweight RepYOLOv5m detector with a dual-stream SlowFast-3D recognition branch, enabling automated inference from raw video input to behavior labels. To improve robustness in dense and occluded scenes, the front-end detector serves as a spatial-prior module, while a decoupled training strategy reduces the impact of localization instability on back-end spatiotemporal learning. In addition, two task-oriented modules are introduced in the recognition branch: the Spatiotemporal Depthwise-Separable 3D module (SDS3D) and the Normalization-Based Temporal Attention Mechanism (NTAM). Experimental results on a real classroom dataset show that RepYOLOv5-SF3D achieves a mean average precision (mAP) of 88.83%, outperforming the baseline SlowFast model by 3.36% and surpassing the existing LSTC method by 2.05%, while maintaining a front-end inference latency of 12.5 ms per frame and a total model size of 151.46 MB. These results demonstrate a favorable balance between fine-grained recognition accuracy and edge-deployment efficiency in practical classroom visual sensing.