MS-YOLO: a multi-scale model for accurate and efficient blood cell detection

Simple SummaryWe propose a new model, YOLOv5-SA-FC, for efficient pig population detection and counting in intelligent breeding. Traditional manual methods are slow and inaccurate. Our model incorporates shuffle attention (SA) and Focal-CIoU (FC) for an improved performance. SA enhances feature extraction without adding parameters, and FC reduces the sample imbalance impact. Our experiments show that YOLOv5-SA-FC achieves a 93.8% mean average precision (mAP) and 95.6% count accuracy, outperforming other methods by 10.2% and 15.8% in pig detection and counting. This validates its effectiveness in intelligent pig breeding.The efficient detection and counting of pig populations is critical for the promotion of intelligent breeding. Traditional methods for pig detection and counting mainly rely on manual labor, which is either time-consuming and inefficient or lacks sufficient detection accuracy. To address these issues, a novel model for pig detection and counting based on YOLOv5 enhanced with shuffle attention (SA) and Focal-CIoU (FC) is proposed in this paper, which we call YOLOv5-SA-FC. The SA attention module in this model enables multi-channel information fusion with almost no additional parameters, enhancing the richness and robustness of feature extraction. Furthermore, the Focal-CIoU localization loss helps to reduce the impact of sample imbalance on the detection results, improving the overall performance of the model. From the experimental results, the proposed YOLOv5-SA-FC model achieved a mean average precision (mAP) and count accuracy of 93.8% and 95.6%, outperforming other methods in terms of pig detection and counting by 10.2% and 15.8%, respectively. These findings verify the effectiveness of the proposed YOLOv5-SA-FC model for pig population detection and counting in the context of intelligent pig breeding.

An ultrasound standard plane detection model of fetal head based on multi-task learning and hybrid knowledge graph

Object detection is an important process in surveillance system to locate objects and it is considered as major application in computer vision. The Convolution Neural Network (CNN) based models have been developed by many researchers for object detection to achieve higher performance. However, existing models have some limitations such as overfitting problem and lower efficiency in small object detection. Object detection in remote sensing hasthe limitations of low efficiency in detecting small object and the existing methods have poor localization. Cascade Object Detection methods have been applied to increase the learning process of the detection model. In this research, the Additive Activation Function (AAF) is applied in a Faster Region based CNN (RCNN) for object detection. The proposed AAF-Faster RCNN method has the advantage of better convergence and clear bounding variance. The Fourier Series and Linear Combination of activation function are used to update the loss function. The Microsoft (MS) COCO datasets and Pascal VOC 2007/2012 are used to evaluate the performance of the AAF-Faster RCNN model. The proposed AAF-Faster RCNN is also analyzed for small object detection in the benchmark dataset. The analysis shows that the proposed AAF-Faster RCNN model has higher efficiency than state-of-art Pay Attention to Them (PAT) model in object detection. To evaluate the performance of AAF-Faster RCNN method of object detection in remote sensing, the NWPU VHR-10 remote sensing data set is used to test the proposed method. The AAF-Faster RCNN model has mean Average Precision (mAP) of 83.1% and existing PAT-SSD512 method has the 81.7%mAP in Pascal VOC 2007 dataset.

When the number of white blood cells (WBCs) in the human body becomes imbalanced, leukemia (blood cancer) can develop, affecting individuals of all ages. Early and accurate detection is crucial for improving patient outcomes, yet manual diagnosis is time-consuming and prone to subjectivity. This study proposes an efficient hybrid convolutional neural network (CNN)-transformer model for automated blood cancer detection, integrating convolutional layers for localized feature extraction with transformer-based attention mechanisms for capturing global dependencies. The architecture employs depthwise separable convolutions, efficient multi-head self-attention (EMHSA), and an efficient multilayer perceptron (EMLP) block, optimized via Bayesian hyperparameter tuning. The proposed model was evaluated on two publicly available datasets: the Blood Cancer dataset (binary classification) and the Blood Cells Cancer (ALL) dataset (four-class classification). Using a 50:50 training-testing split, the model achieved 100% accuracy, 100% precision, 100% recall, and 100% F1-score on the Blood Cancer dataset, and 99% accuracy, 99% precision, 100% recall, and 99% F1-score on the ALL dataset. Additional experiments with 80:20, 70:30, and 60:40 splits confirmed consistent performance above 98% across all metrics, demonstrating strong robustness. The model contains only 2.04 million trainable parameters, significantly fewer than standard CNN or transformer-based architectures, making it computationally lightweight and suitable for deployment in resource-constrained clinical environments. These results highlight the potential of the proposed hybrid framework to provide accurate and efficient blood cancer classification, advancing the applicability of deep learning in hematological diagnostics.

Psychometric methods for accurate and timely detection of item compromise have been a long‐standing topic. While Bayesian methods can incorporate prior knowledge or expert inputs as additional information for item compromise detection, they have not been employed in item compromise detection itself. The current study proposes a two‐phase Bayesian change‐point framework for both stationary and real‐time detection of changes in each item's compromise status. In Phase I, a stationary Bayesian change‐point model for compromise detection is fitted to the observed responses over a specified time‐frame. The model produces parameter estimates for the change‐point of each item from uncompromised to compromised, as well as structural parameters accounting for the post‐change response distribution. Using the post‐change model identified in Phase I, the Shiryaev procedure for sequential testing is employed in Phase II for real‐time monitoring of item compromise. The proposed methods are evaluated in terms of parameter recovery, detection accuracy, and detection efficiency under various simulation conditions and in a real data example. The proposed method also showed superior detection accuracy and efficiency compared to the cumulative sum procedure.

Objective. Sleep apnea is a common sleep disorder associated with severe health risks, necessitating accurate and efficient detection methods.Approach. This study proposes ModelS4Apnea, a deep learning framework for sleep apnea detection from electrocardiogram (ECG) spectrograms, integrating structured state space models (S4) for temporal modeling. The framework consists of a convolutional neural network module for local feature extraction, an S4 module for capturing long-range dependencies, and a classification module for final predictions.Main results. The model was trained and evaluated on the Apnea-ECG dataset, achieving an accuracy of 0.933, anF1-score of 0.912, a sensitivity of 0.916, and a specificity of 0.944, outperforming most prior studies while maintaining computational efficiency.Significance. Compared to existing methods, ModelS4Apnea provides high classification performance with significantly fewer trainable parameters than long short-term memory-based models, reducing training time and memory consumption. The model's ability to aggregate segment-level predictions enabled perfect per-recording classification, demonstrating its robustness in diagnosing sleep apnea across entire recordings. Moreover, its low memory footprint and fast inference speed make it well-suited for wearable devices, home-based monitoring, and clinical applications, offering a scalable and efficient solution for automated sleep apnea detection. Future work may explore multi-modal data integration, real-world deployment, and further optimizations to enhance its clinical applicability and reliability.

Deep learning technology can automatically learn features from large amounts of data, with powerful feature extraction and pattern recognition capabilities, thereby improving the accuracy and efficiency of object detection. [The objective of this study]: In order to improve the accuracy and speed of mask wearing deep learning detection models in the post pandemic era, the [Problem this study aimed to resolve] was based on the fact that no research work has been reported on standardized detection models for mask wearing with detecting nose targets specially. [The topic and method of this study]: A mask wearing normalization detection model (towards the wearing style exposing the nose to outside, which is the most obvious characteristic of non-normalized style) based on improved YOLOv5s (You Only Look Once v5s is an object detection network model) was proposed. [The improved method of the proposed model]: The improvement design work of the detection model mainly includes (1) the BottleneckCSP (abbreviation of Bottleneck Cross Stage Partial) module was improved to a BottleneckCSP-MASK (abbreviation of Bottleneck Cross Stage Partial-MASK) module, which was utilized to replace the BottleneckCSP module in the backbone architecture of the original YOLOv5s model, which reduced the weight parameters' number of the YOLOv5s model while ensuring the feature extraction effect of the bonding fusion module. (2) An SE module was inserted into the proposed improved model, and the bonding fusion layer in the original YOLOv5s model was improved for better extraction of the features of mask and nose targets. [Results and validation]: The experimental results indicated that, towards different people and complex backgrounds, the proposed mask wearing normalization detection model can effectively detect whether people are wearing masks and whether they are wearing masks in a normalized manner. The overall detection accuracy was 99.3% and the average detection speed was 0.014 s/pic. Contrasted with original YOLOv5s, v5m, and v5l models, the detection results for two types of target objects on the test set indicated that the mAP of the improved model increased by 0.5%, 0.49%, and 0.52%, respectively, and the size of the proposed model compressed by 10% compared to original v5s model. The designed model can achieve precise identification for mask wearing behaviors of people, including not wearing a mask, normalized wearing, and wearing a mask non-normalized.

In this study, we present a novel approach leveraging the segment anything model (SAM) for the efficient detection and tracking of vehicles in urban traffic surveillance systems by utilizing uncalibrated low-resolution highway cameras. This research addresses the critical need for accurate vehicle monitoring in intelligent transportation systems (ITS) and smart city infrastructure. Traditional methods often struggle with the variability and complexity of urban environments, leading to suboptimal performance. Our approach harnesses the power of SAM, an advanced deep learning-based image segmentation algorithm, to significantly enhance the detection accuracy and tracking robustness. Through extensive testing and evaluation on two datasets of 511 highway cameras from Quebec, Canada and NVIDIA AI City Challenge Track 1, our algorithm achieved exceptional performance metrics including a precision of 89.68%, a recall of 97.87%, and an F1-score of 93.60%. These results represent a substantial improvement over existing state-of-the-art methods such as the YOLO version 8 algorithm, single shot detector (SSD), region-based convolutional neural network (RCNN). This advancement not only highlights the potential of SAM in real-time vehicle detection and tracking applications, but also underscores its capability to handle the diverse and dynamic conditions of urban traffic scenes. The implementation of this technology can lead to improved traffic management, reduced congestion, and enhanced urban mobility, making it a valuable tool for modern smart cities. The outcomes of this research pave the way for future advancements in remote sensing and photogrammetry, particularly in the realm of urban traffic surveillance and management.

Label-free cell separation and sorting in a microfluidic system, an essential technique for modern cancer diagnosis, resulted in high-throughput single-cell analysis becoming a reality. However, designing an efficient cell detection model is challenging. Traditional cell detection methods are subject to occlusion boundaries and weak textures, resulting in poor performance. Modern detection models based on convolutional neural networks (CNNs) have achieved promising results at the cost of a large number of both parameters and floating point operations (FLOPs). In this work, we present a lightweight, yet powerful cell detection model named CellNet, which includes two efficient modules, CellConv blocks and the h-swish nonlinearity function. CellConv is proposed as an effective feature extractor as a substitute to computationally expensive convolutional layers, whereas the h-swish function is introduced to increase the nonlinearity of the compact model. To boost the prediction and localization ability of the detection model, we re-designed the model’s multi-task loss function. In comparison with other efficient object detection methods, our approach achieved state-of-the-art 98.70% mean average precision (mAP) on our custom sea urchin embryos dataset with only 0.08 M parameters and 0.10 B FLOPs, reducing the size of the model by 39.5× and the computational cost by 4.6×. We deployed CellNet on different platforms to verify its efficiency. The inference speed on a graphics processing unit (GPU) was 500.0 fps compared with 87.7 fps on a CPU. Additionally, CellNet is 769.5-times smaller and 420 fps faster than YOLOv3. Extensive experimental results demonstrate that CellNet can achieve an excellent efficiency/accuracy trade-off on resource-constrained platforms.

Backscatter has significant influence on detection efficiency for underwater lidar, especially for coaxial photon-counting lidar using a Geiger-mode avalanche photodiode. In this paper, based on our underwater coaxial photon-counting lidar structure and volume scatter function, a detection model with consideration of backscatter and refraction indices is proposed. Using this detection model, analysis of the detection efficiency is conducted. It reveals that in an underwater environment, higher pulse energy or a closer target range is not necessarily helpful for a higher target detection probability, which is vastly different from our traditional concepts. For example, the detection probability for a 5 m target would be 0.76 using a 200 pJ pulse and 0.55 using a 1000 pJ pulse for our coaxial photon-counting lidar. Monte Carlo simulation is conducted to verify our model and analysis, and some practical methods for improving the target detection probability in an underwater environment are proposed.

Malware detection is a vital problem, and efficient methods that can efficiently detect malware are needed. The increasing use of mobile computers makes malware detection a vital part of security in an era where smartphones have come to play a key role in many of our daily lives. Earlier approaches, however, suffer from high false positive rates; they are not scalable for larger databases, or they are not amenable to adapt well to novel zero-day malware. For these reasons, the demand for more sensitive and flexible detection models is high. In this study, we develop a hybrid mobile malware detection framework that leverages ant colony optimization (ACO) and deep neural networks (DNNs) to improve detection accuracy, reduce the rate of false positives, and make the model resilient to new malware. AntDroidNet is a novel ACO-enabled feature selection model that dynamically reduces the feature dimensionality by selecting single instances to include the most informative properties and avoid dimensionality. A DNN is consequently constructed to train the determined set of features, improving the identified classification performance and decreasing the number of instances with false discoveries. In this way, a self-optimizing feedback loop can iteratively improve the feature selection process given the performance of the DNN, leading to a dynamic and efficient detection model. Using the CICMalDroid2020 dataset, the proposed AntDroidNet model achieves a remarkable accuracy of 99.89% and an excellent false positive rate of only 0.13% and outperforms the classical machine learning algorithms in terms of accuracy and efficiency. AntDroidNet is a scalable and powerful mobile malware detection model that eclipses all state-of-the-art methods and shows important enhancements in efficiency and reliability. By prototyping whitelisting systems, this work opens new avenues in mobile security and lays the groundwork for future work on building real-time detection components and system components able to scale to the fast pace of evolution of mobile malware in new connected ecosystems.

EFDet: An efficient detection method for cucumber disease under natural complex environments

The widespread use of unmanned aerial vehicles (UAVs) has brought many benefits, particularly for military and civil applications. For example, UAVs can be used in communication, ecological surveys, agriculture, and logistics to improve efficiency and reduce the required workforce. However, the malicious use of UAVs can significantly endanger public safety and pose many challenges to society. Therefore, detecting malicious UAVs is an important and urgent issue that needs to be addressed. In this study, a combined UAV detection model (CUDM) based on analyzing video abnormal behavior is proposed. CUDM uses abnormal behavior detection models to improve the traditional object detection process. The work of CUDM can be divided into two stages. In the first stage, our model cuts the video into images and uses the abnormal behavior detection model to remove a large number of useless images, improving the efficiency and real-time detection of suspicious targets. In the second stage, CUDM works to identify whether the suspicious target is a UAV or not. Besides, CUDM relies only on ordinary equipment such as surveillance cameras, avoiding the use of expensive equipment such as radars. A self-made UAV dataset was constructed to verify the reliability of CUDM. The results show that CUDM not only maintains the same accuracy as state-of-the-art object detection models but also reduces the workload by 32%. Moreover, it can detect malicious UAVs in real-time.

In this work, we propose an optimal solution for face detection when operating in the thermal and visible bands. Our aim is to train, fine tune, optimize and validate preexisting object detection models using thermal and visible data separately. Thus, we perform an empirical study to determine the most efficient band specific DeepFace detection model in terms of detection performance. The original object detection models that were selected for our study are the Faster R-CNN (Region based Convolutional Neural Network), SSD (Single-shot Multi-Box Detector) and R-FCN (Region-based Fully Convolutional Network). Also, the dual-band dataset used for this work is composed of two challenging MWIR and visible band face datasets, where the faces were captured under variable conditions, i.e. indoors, outdoors, different standoff distances (5 and 10 meters) and poses. Experimental results show that the proposed detection model yields the highest accuracy independent of the band and scenario used. Specifically, we show that a modified and tuned Faster R-CNN architecture with ResNet 101 is the most promising model when compared to all the other models tested. The proposed model yields accuracy of 99.2% and 98.4% when tested on thermal and visible face data respectively. Finally, while the proposed model is relatively slower than its competitors, our further experiments show that the speed of this network can be increased by reducing the number of proposals in RPN (Region Proposal Network), and thus, the computational complexity challenge is significantly minimized.

With the aim of solving the detection problems for current complex event detection models in detecting a related event for a complex event from the high proportion disordered RFID event stream due to its big uncertainty arrival, an efficient complex event detection model based on Extended Nondeterministic Finite Automaton (ENFA) is proposed in this paper. The achievement of the paper rests on the fact that an efficient complex event detection model based on ENFA is presented to successfully realize the detection of a related event for a complex event from the high proportion disordered RFID event stream. Specially, in our model, we successfully use a new ENFA-based complex event detection model instead of an NFA-based complex event detection model to realize the detection of the related events for a complex event from the high proportion disordered RFID event stream by expanding the traditional NFA-based detection model, which can effectively address the problems above. The experimental results show that the proposed model in this paper outperforms some general models in saving detection time, memory consumption, detection latency and improving detection throughput for detecting a related event of a complex event from the high proportion out-of-order RFID event stream.