TI-YOLO: A Lightweight and Efficient Anatomical Structure Detection Model for Tracheal Intubation.

Deploying target detection models on edge devices such as UAVs is challenging due to their limited size and computational capacity, while target detection models typically require significant computational resources. To address this issue, this study proposes a lightweight real-time infrared object detection model named LRI-YOLO (Lightweight Real-time Infrared YOLO), which is based on YOLOv8n. The model improves the C2f module’s Bottleneck structure by integrating Partial Convolution (PConv) with Pointwise Convolution (PWConv), achieving a more lightweight design. Furthermore, during the feature fusion stage, the original downsampling structure with ordinary convolution is replaced with a combination of max pooling and regular convolution. This modification retains more feature map information. The model’s structure is further optimized by redesigning the decoupled detection head with Group Convolution (GConv) instead of ordinary convolution, significantly enhancing detection speed. Additionally, the original BCELoss is replaced with EMASlideLoss, a newly developed classification loss function introduced in this study. This loss function allows the model to focus more on hard samples, thereby improving its classification capability. Compared to the YOLOv8n algorithm, LRI-YOLO is more lightweight, with its parameters reduced by 46.7% and floating-point operations (FLOPs) reduced by 53.1%. Moreover, the mean average precision (mAP) reached 94.1%. Notably, on devices with moderate computational power that only have a Central Processing Unit (CPU), the detection speed reached 42 frames per second (FPS), surpassing most mainstream models. This indicates that LRI-YOLO offers a novel solution for real-time infrared object detection on edge devices such as drones.

With the rapid development of autonomous driving technology, traditional object detection algorithms often face challenges such as the trade-off between model parameters and recognition accuracy, insufficient detection accuracy, and slow detection speed when dealing with vehicle and pedestrian detection tasks in complex scenarios. To address these challenges, we propose a novel lightweight real-time driving scene detection network, FEEP-YOLO. Based on the YOLOv8 algorithm, our network not only maintains high detection accuracy but also achieves faster training speed and fewer model parameters. FEEP-YOLO employs a lightweight network, Faster Net, to significantly reduce model parameters, thereby accelerating detection speed and better meeting real-time requirements. To further improve detection accuracy, we introduce an EMA attention module into the network and optimize the bounding box loss function, adopting the EIOU loss function to enhance target features and reduce background interference. Additionally, to detect small objects more accurately, we add feature layers rich in semantic information to the detection end. Experimental results demonstrate that FEEP-YOLO outperforms YOLOv8 in terms of both detection speed and accuracy on the KITTI and SODA datasets, while significantly reducing the number of model parameters and model size. FEEP-YOLO achieves high detection speed while improving detection accuracy and reducing computational resource consumption, meeting the practical application requirements under limited computing power, memory space and power consumption. The key contribution of FEEP-YOLO lies in its ability to balance accuracy speed, and efficiency, making it a robust solution for real-time autonomous driving applications in complex environments.

Stress is a common human reaction to demanding circumstances, and prolonged and excessive stress can have detrimental effects on both mental and physical health. Heart rate variability (HRV) is widely used as a measure of stress due to its ability to capture variations in the time intervals between heartbeats. However, achieving high accuracy in stress detection through machine learning (ML), using a reduced set of statistical features extracted from HRV, remains a significant challenge. In this study, we aim to address these challenges by proposing lightweight ML models that can effectively detect stress using minimal HRV features and are computationally efficient enough for IoT deployment. We have developed ML models incorporating efficient feature selection techniques and hyper-parameter tuning. The publicly available SWELL-KW dataset has been utilized for evaluating the performance of our models. Our results demonstrate that lightweight models such as -NN and Decision Tree can achieve competitive accuracy while ensuring lower computational demands, making them ideal for real-time applications. Promisingly, among the developed models, the -nearest neighbors (-NN) algorithm has emerged as the best-performing model, achieving an accuracy score of 99.3 using only three selected features. To confirm real-world deployability, we benchmarked the best model on an 8 GB NVIDIA Jetson Orin Nano edge device, where it retained 99.26% accuracy and completed training in 31 s. Furthermore, our study has incorporated local interpretable model-agnostic explanations to provide comprehensive insights into the predictions made by the -NN-based architecture.

Forest fires pose significant hazards to ecological environments and economic society. The detection of forest fire smoke can provide crucial information for the suppression of early fires. Previous detection models based on deep learning have been limited in detecting small smoke and smoke with smoke-like interference. In this paper, we propose a lightweight model for forest fire smoke detection that is suitable for UAVs. Firstly, a smoke dataset is created from a combination of forest smoke photos obtained through web crawling and enhanced photos generated by using the method of synthesizing smoke. Secondly, the GSELAN and GSSPPFCSPC modules are built based on Ghost Shuffle Convolution (GSConv), which efficiently reduces the number of parameters in the model and accelerates its convergence speed. Next, to address the problem of indistinguishable feature boundaries between clouds and smoke, we integrate coordinate attention (CA) into the YOLO feature extraction network to strengthen the extraction of smoke features and attenuate the background information. Additionally, we use Content-Aware Reassembly of FEatures (CARAFE) upsampling to expand the receptive field in the feature fusion network and fully exploit the semantic information. Finally, we adopt SCYLLA-Intersection over Union (SIoU) loss as a replacement for the original loss function in the prediction phase. This substitution leads to improved convergence efficiency and faster convergence. The experimental results demonstrate that the LMDFS model proposed for smoke detection achieves an accuracy of 80.2% with a 5.9% improvement compared to the baseline and a high number of Frames Per Second (FPS)—63.4. The model also reduces the parameter count by 14% and Giga FLoating-point Operations Per second (GFLOPs) by 6%. These results suggest that the proposed model can achieve a high accuracy while requiring fewer computational resources, making it a promising approach for practical deployment in applications for detecting smoke.

A detection method for dead caged hens based on improved YOLOv7

The utilization of Synthetic Aperture Radar (SAR) for real-time ship detection proves highly advantageous in the supervision and monitoring of maritime activities. Ship detection in the range-compressed domain of SAR rather than in fully focused SAR imagery can significantly reduce the time and computational resources required for complete SAR imaging, enabling lightweight real-time ship detection methods to be implemented on an airborne or spaceborne SAR platform. However, there is a lack of lightweight ship detection methods specifically designed for the SAR range-compressed domain. In this paper, we propose Fast Range-Compressed Detection (FastRCDet), a novel lightweight network for ship detection in the SAR range-compressed domain. Firstly, to address the distinctive geometric characteristics of the SAR range-compressed domain, we propose a Lightweight Adaptive Network (LANet) as the backbone of the network. We introduce Arbitrary Kernel Convolution (AKConv) as a fundamental component, which enables the flexible adjustment of the receptive field shape and better adaptation to the large scale and aspect ratio characteristics of ships in the range-compressed domain. Secondly, to enhance the efficiency and simplicity of the network model further, we propose an innovative Multi-Scale Fusion Head (MSFH) module directly integrated after the backbone, eliminating the need for a neck module. This module effectively integrates features at various scales to more accurately capture detailed information about the target. Thirdly, to further enhance the network’s adaptability to ships in the range-compressed domain, we propose a novel Direction IoU (DIoU) loss function that leverages angle cost to control the convergence direction of predicted bounding boxes, thereby improving detection accuracy. Experimental results on a publicly available dataset demonstrate that FastRCDet achieves significant reductions in parameters and computational complexity compared to mainstream networks without compromising detection performance in SAR range-compressed images. FastRCDet achieves a low parameter of 2.49 M and a high detection speed of 38.02 frames per second (FPS), surpassing existing lightweight detection methods in terms of both model size and processing rate. Simultaneously, it attains an average accuracy (AP) of 77.12% in terms of its detection performance. This method provides a baseline in lightweight network design for SAR ship detection in the range-compressed domain and offers practical implications for resource-constrained embedded platforms.

A challenging and attractive task in computer vision is underwater object detection. Although object detection techniques have achieved good performance in general datasets, problems of low visibility and color bias in the complex underwater environment have led to generally poor image quality; besides this, problems with small targets and target aggregation have led to less extractable information, which makes it difficult to achieve satisfactory results. In past research of underwater object detection based on deep learning, most studies have mainly focused on improving detection accuracy by using large networks; the problem of marine underwater lightweight object detection has rarely gotten attention, which has resulted in a large model size and slow detection speed; as such the application of object detection technologies under marine environments needs better real-time and lightweight performance. In view of this, a lightweight underwater object detection method based on the MobileNet v2, You Only Look Once (YOLO) v4 algorithm and attentional feature fusion has been proposed to address this problem, to produce a harmonious balance between accuracy and speediness for target detection in marine environments. In our work, a combination of MobileNet v2 and depth-wise separable convolution is proposed to reduce the number of model parameters and the size of the model. The Modified Attentional Feature Fusion (AFFM) module aims to better fuse semantic and scale-inconsistent features and to improve accuracy. Experiments indicate that the proposed method obtained a mean average precision (mAP) of 81.67% and 92.65% on the PASCAL VOC dataset and the brackish dataset, respectively, and reached a processing speed of 44.22 frame per second (FPS) on the brackish dataset. Moreover, the number of model parameters and the model size were compressed to 16.76% and 19.53% of YOLO v4, respectively, which achieved a good tradeoff between time and accuracy for underwater object detection.

With the increasing complexity of the shortwave communication environment, the efficiency and accuracy of the manual detection of Morse code no longer meet actual needs. Therefore, this paper proposes a Morse code detection algorithm called YFDM. For the time–frequency image of the received signal, a combination module of deformable convolution and C3 is used to enhance the backbone network’s attention to the abstract semantics and location information of Morse code. GSConv and VOV-GSCSP modules are used to build a lightweight neck network. Finally, the confidence propagation cluster (CP-Cluster) algorithm is used to filter the detection frame. In an ablation experiment, the parameters and giga floating-point operations per second (GFLOPs) of YFDM were 5.961 M and 9.74 G, respectively, 15.11% and 38.9% less than those of YOLOv5. Moreover, when WIoUv1 was used as the loss function of the bounding box, the AP0.5:0.95 and frames per second (FPS) values of the algorithm reached the highest values, 0.68 and 72.4. The experimental results indicate that the algorithm can effectively reduce the weight of the model while ensuring the detection accuracy and inference speed.

Lightweight object detection and recognition models are extremely crucial for in-orbit applications, which is the most critical factor for whether deep learning-based object detection and recognition algorithms can be applied to remote sensing satellites for real-time or near real-time processing. Global information is extremely important for object detection and recognition of remote sensing images. However, due to the high computational cost, the existing CNN-based lightweight models over-emphasize on the extraction of local information, while ignoring the global information. For this reason, we propose a lightweight object detection and recognition model (Lightweight Global-Local Detection, LGLDet) based on the especially light global modeling structure. In LGLDet, light global-local module (LGLM) is proposed to extract the global and local information. The LGLM consists of Point2Patch Non-Local (P2PNL), local branch and skip connection. Specifically, P2PNL is proposed to reduce the computation of global long-range dependency modeling. In addition, the feature fusion part and detection head are also designed in a lightweight way. In the experiments, the proposed method can achieve optimal performance with fewer parameters and lower computational complexity than existing CNN-based lightweight models and transformer-based lightweight models with similar parameters or computational complexity. The code will be released on the site of https://github.com/dyl96/LGLDet.

In the field of forest fire detection, accurately identifying smoke and flames has always posed a technical challenge. This study addresses this issue by proposing an improved model based on YOLOv12. The model adopts lightweight design principles and feature enhancement techniques, significantly improving detection accuracy and efficiency, and providing solid technical support for early warning of forest fires. After selecting YOLOv12 as the base architecture, the research team further integrated the AKConv variable kernel convolution module and the CBAM attention module, and incorporated the BiFPN weighted bidirectional feature pyramid network. The combination of these technologies aims to optimize the feature fusion process and significantly enhance the model’s ability to detect small targets. The improved model achieved an average precision (mAP) of 72.8% on a self-built dataset, with a 7% increase in detection speed compared to the original model, and its robustness in complex scenarios has been significantly enhanced. This strategy not only improves detection accuracy through dynamic feature enhancement and efficient cross-scale fusion techniques, but also effectively reduces redundancy in the computation process, making it possible to build an “air-ground-space” integrated monitoring network, and promoting the shift towards intelligent early warning in forest fire prevention efforts.

Simple SummaryThe aim of this study was to improve the hatching efficiency of duck eggs by automatically assessing the hatching characteristics of early hatching eggs. The timely and accurate detection of fertile and infertile eggs is an important research topic in the breeder egg hatching industry. Detecting infertile eggs early in the hatching process not only improves the hatching efficiency of duck eggs, but also brings benefits to hatching companies. In recent years, the rapid development of deep learning and computer vision technology has brought us new ideas. We propose a lightweight multi-target detection method based on deep learning to evaluate the hatching characteristics of duck eggs. The results show that the method could meet the requirements for accuracy and real-time detection in industrial production.Since it is difficult to accurately identify the fertilization and infertility status of multiple duck eggs on an incubation tray, and due to the lack of easy-to-deploy detection models, a novel lightweight detection architecture (LDA) based on the YOLOX-Tiny framework is proposed in this paper to identify sterile duck eggs with the aim of reducing model deployment requirements and improving detection accuracy. Specifically, the method acquires duck egg images through an acquisition device and augments the dataset using rotation, symmetry, and contrast enhancement methods. Then, the traditional convolution is replaced by a depth-wise separable convolution with a smaller number of parameters, while a new CSP structure and backbone network structure are used to reduce the number of parameters of the model. Finally, to improve the accuracy of the network, the method includes an attention mechanism after the backbone network and uses the cosine annealing algorithm in training. An experiment was conducted on 2111 duck eggs, and 6488 duck egg images were obtained after data augmentation. In the test set of 326 duck egg images, the mean average precision (mAP) of the method in this paper was 99.74%, which was better than the 94.92% of the YOLOX-Tiny network before improvement, and better than the reported prediction accuracy of 92.06%. The number of model parameters was only 1.93 M, which was better than the 5.03 M of the YOLOX-Tiny network. Further, by analyzing the concurrent detection of single 3 × 5, 5 × 7 and 7 × 9 grids, the algorithm achieved a single detection number of 7 × 9 = 63 eggs. The method proposed in this paper significantly improves the efficiency and detection accuracy of single-step detection of breeder duck eggs, reduces the network size, and provides a suitable method for identifying sterile duck eggs on hatching egg trays. Therefore, the method has good application prospects.

Current infusion monitoring methods primarily rely on two technological approaches: non-visual sensor technology and visual sensor technology, for real-time monitoring of the remaining liquid volume in infusion bottles within infusion scenarios. However, non-visual sensor-based methods often suffer from complex installation procedures and are prone to external interference, while visual sensor-based methods tend to exhibit low detection accuracy in complex infusion environments involving small targets, low contrast, tilted objects, and partial occlusions, making it difficult to accurately monitor the remaining liquid. To address these challenges, we propose a high-precision and lightweight object detection algorithm-SDM-YOLO11n-based on an improved version of YOLO11n. Specifically, a lightweight Split-Parallel Convolution (SPConv) module is introduced to enhance the backbone network's spatial modeling capabilities and improve feature extraction efficiency; the traditional upsampling operation is replaced with a Dynamic Sampling (DySample) module for more adaptive feature reconstruction and multi-scale information fusion; and a Mixed Local Channel Attention (MLCA) mechanism is incorporated to strengthen attention to key regions of infusion bottles and their internal liquids, thereby further improving detection accuracy. In addition, a method based on the ratio of geometric parameters of oriented bounding boxes is proposed to precisely estimate the remaining liquid volume in infusion bottles. Experimental results show that SDM-YOLO11n improves mAP@0.5:0.95 by 0.6 percentage points compared to YOLO11n, with a model size of only 5.1 MB. The proposed algorithm achieves high-precision detection of infusion bottles and their internal liquids in complex scenarios and enables real-time monitoring of the remaining liquid volume in multiple infusion bottles.

YOLOv5s-road: Road surface defect detection under engineering environments based on CNN-transformer and adaptively spatial feature fusion

The fuel injection system, as a crucial component of engineering vehicles, can lead to a decline in vehicle performance or even safety accidents when it malfunctions. Detection of fuel injection pulse signals is a key technique for diagnosing faults in the fuel injection system. Traditional fault detection methods are inefficient and struggle to cope with complex industrial environments. Meanwhile, deep learning-based network models require substantial computational resources, making them difficult to deploy on embedded devices. To address these issues, this paper proposes an improved lightweight fuel injection system fault detection model based on GhostNet. This model enhances the Ghost module, incorporates SE (Squeeze-and-Excitation) attention to improve accuracy, and replaces some of the Ghost modules with depthwise separable convolution to reduce computational resource consumption. Validation was performed on a self-constructed dataset of fuel injection pulse signals, consisting of 15,000 samples from normal and four fault conditions. Compared with seven lightweight neural networks, the proposed model achieved an average accuracy reduction of 1.42 %, while reducing model parameters by 98.99 %, floating-point operations by <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{9 5. 7 9 \%}$</tex>, and model size by <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{9 6. 5 4 \%}$</tex>. Deployed on the Arduino Nano 33 BLE Sense, the model occupied 46.58 % of Flash memory and 38.44 % of SRAM, with an inference latency of <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{2 6 1 ~ m s}$</tex>. This study demonstrates the model's high detection accuracy and efficiency on embedded devices, highlighting its significant engineering value for enhancing fuel injection system reliability and driving safety, and providing a valuable reference for practical applications of deep learning models in fault detection of engineering vehicle fuel injection systems.

An accurate detection and segmentation model of obscured green fruits