Channel Shuffle Research Articles

CSSANet: A Channel Shuffle Slice-Aware Network for Pulmonary Nodule Detection

Lung cancer stands as the leading cause of cancer related mortality worldwide. Precise and automated identification of lung nodules through 3D Computed Tomography (CT) scans is an essential part of screening for lung cancer effectively. Due to the small size of pulmonary nodules and the close correlation between neighboring slices of 3D CT images, most of the existing methods only consider the characteristics of a single slice, thus easily lead to insufficient detection accuracy of pulmonary nodules. To solve this problem, this paper proposes a Channel Shuffle Slice-Aware Network (CSSANet), which aims to fully exploit the spatial correlation between slices and effectively utilize the intra-slice features and inter-slice contextual information to achieve accurate detection of lung nodules. Specifically, we design a Group Shuffle Attention module (GSA module) to fuse the inter-slice feature in order to enhance the discrimination and extraction of corresponding shape information of distinct nodules in the same group of slices. Experiments and ablation study on a publicly available LUNA16 dataset demonstrate that the proposed method can enhance the detection sensitivity effectively. The Competition Performance Metric (CPM) score of 89.8% is superior over other representative detection models.

Flame and smoke detection based on channel shuffling and adaptive spatial feature fusion

Flame and smoke detection based on channel shuffling and adaptive spatial feature fusion

WGS-YOLO: A real-time object detector based on YOLO framework for autonomous driving

The safety and reliability of autonomous driving depends on the precision and efficiency of object detection systems. In this paper, a refined adaptation of the YOLO architecture (WGS-YOLO) is developed to improve the detection of pedestrians and vehicles. Specifically, its information fusion is enhanced by incorporating the Weighted Efficient Layer Aggregation Network (W-ELAN) module, an innovative dynamic weighted feature fusion module using channel shuffling. Meanwhile, the computational demands and parameters of the proposed WGS-YOLO are significantly reduced by employing the Space-to-Depth Convolution (SPD-Conv) and the Grouped Spatial Pyramid Pooling (GSPP) modules that have been strategically designed. The performance of our model is evaluated with the BDD100k and DAIR-V2X-V datasets. In terms of mean Average Precision (mAP0.5), the proposed model outperforms the baseline Yolov7 by 12%. Furthermore, extensive experiments are conducted to verify our analysis and the model’s robustness across diverse scenarios.

Fault Diagnosis Method for Vacuum Contactor Based on Time-Frequency Graph Optimization Technique and ShuffleNetV2.

This paper presents a fault diagnosis method for a vacuum contactor using the generalized Stockwell transform (GST) of vibration signals. The objective is to solve the problem of low diagnostic performance efficiency caused by the inadequate feature extraction capability and the redundant pixels in the graph background. The proposed method is based on the time-frequency graph optimization technique and ShuffleNetV2 network. Firstly, vibration signals in different states are collected and converted into GST time-frequency graphs. Secondly, multi-resolution GST time-frequency graphs are generated to cover signal characteristics in all frequency bands by adjusting the GST Gaussian window width factor λ. The OTSU algorithm is then combined to crop the energy concentration area, and the size of these time-frequency graphs is optimized by 68.86%. Finally, considering the advantages of the channel split and channel shuffle methods, the ShuffleNetV2 network is adopted to improve the feature learning ability and identify fault categories. In this paper, the CKJ5-400/1140 vacuum contactor is taken as the test object. The fault recognition accuracy reaches 99.74%, and the single iteration time of model training is reduced by 19.42%.

DA-YOLOv7: A Deep Learning-Driven High-Performance Underwater Sonar Image Target Recognition Model

Affected by the complex underwater environment and the limitations of low-resolution sonar image data and small sample sizes, traditional image recognition algorithms have difficulties achieving accurate sonar image recognition. The research builds on YOLOv7 and devises an innovative fast recognition model designed explicitly for sonar images, namely the Dual Attention Mechanism YOLOv7 model (DA-YOLOv7), to tackle such challenges. New modules such as the Omni-Directional Convolution Channel Prior Convolutional Attention Efficient Layer Aggregation Network (OA-ELAN), Spatial Pyramid Pooling Channel Shuffling and Pixel-level Convolution Bilat-eral-branch Transformer (SPPCSPCBiFormer), and Ghost-Shuffle Convolution Enhanced Layer Aggregation Network-High performance (G-ELAN-H) are central to its design, which reduce the computational burden and enhance the accuracy in detecting small targets and capturing local features and crucial information. The study adopts transfer learning to deal with the lack of sonar image samples. By pre-training the large-scale Underwater Acoustic Target Detection Dataset (UATD dataset), DA-YOLOV7 obtains initial weights, fine-tuned on the smaller Smaller Common Sonar Target Detection Dataset (SCTD dataset), thereby reducing the risk of overfitting which is commonly encountered in small datasets. The experimental results on the UATD, the Underwater Optical Target Detection Intelligent Algorithm Competition 2021 Dataset (URPC), and SCTD datasets show that DA-YOLOV7 exhibits outstanding performance, with mAP@0.5 scores reaching 89.4%, 89.9%, and 99.15%, respectively. In addition, the model maintains real-time speed while having superior accuracy and recall rates compared to existing mainstream target recognition models. These findings establish the superiority of DA-YOLOV7 in sonar image analysis tasks.

Visualization-based comprehensive feature representation with improved EfficientNet for malicious file and variant recognition

Malicious file attacks seriously affect network and data security, and recognizing malicious files and variants is crucial for preventing network attacks. Faced with the challenge of traditional methods in quickly, effectively, and efficiently recognizing malicious files or variants, visualization-based feature representation methods have shown promising results. However, practical applications encounter issues such as loss of crucial information, high spatiotemporal overhead, and the need for model performance improvement. Therefore, this paper introduces a novel recognition framework focusing on feature representation and model performance. The framework uses the proposed visualization-based comprehensive feature representation method (VCFR) to extract file information into the Gray-Level Co-occurrence Matrix (GLCM), 2-gram frequency matrix, and interval 2-gram frequency matrix, followed by feature fusion to generate the three-channel RGB images. Subsequently, the proposed lightweight model is applied for recognizing those files, which utilizes ideas such as group convolution, channel shuffle, and attention mechanisms to improve model performance while significantly reducing model parameters, size, and FLOPs. In summary, through a series of experiments conducted on manually collected malicious file dataset (MFD) and public dataset MMCC, the proposed framework significantly outperformed other state-of-the-art technologies and has F1-Score as high as 94.10% and 98.58%, respectively, further verifying its outstanding effectiveness and efficiency.

RF-ULM: Ultrasound Localization Microscopy Learned From Radio-Frequency Wavefronts.

In Ultrasound Localization Microscopy (ULM), achieving high-resolution images relies on the precise localization of contrast agent particles across a series of beamformed frames. However, our study uncovers an enormous potential: The process of delay-and-sum beamforming leads to an irreversible reduction of Radio-Frequency (RF) channel data, while its implications for localization remain largely unexplored. The rich contextual information embedded within RF wavefronts, including their hyperbolic shape and phase, offers great promise for guiding Deep Neural Networks (DNNs) in challenging localization scenarios. To fully exploit this data, we propose to directly localize scatterers in RF channel data. Our approach involves a custom super-resolution DNN using learned feature channel shuffling, non-maximum suppression, and a semi-global convolutional block for reliable and accurate wavefront localization. Additionally, we introduce a geometric point transformation that facilitates seamless mapping to the B-mode coordinate space. To understand the impact of beamforming on ULM, we validate the effectiveness of our method by conducting an extensive comparison with State-Of-The-Art (SOTA) techniques. We present the inaugural in vivo results from a wavefront-localizing DNN, highlighting its real-world practicality. Our findings show that RF-ULM bridges the domain shift between synthetic and real datasets, offering a considerable advantage in terms of precision and complexity. To enable the broader research community to benefit from our findings, our code and the associated SOTA methods are made available at https://github.com/hahnec/rf-ulm.

Development of a Technique for Mitigating Four-Wave Mixing Effect in Dense Wavelength Division Multiplexing System Using DP-QPSK with Channel Spacing

A Dense Wavelength Division Multiplexing (DWDM) system is a vital component in the optical network for the transmission of high data over a long distance. However, the non-linear effect of Four-Wave Mixing (FWM), one of the critical non-linear effects, limits the channel capacity and the amount of data that the DWDM system transmits. Intensity Modulation formats and Channel Shuffling algorithms being used to overcome the impact of FWM in DWDM are unable to sufficiently suppress the FWM effect due to limited capacity. This research, therefore, examined the mitigation of the FWM effect in the DWDM system using Dual Polarization Quadrature Phase Shift Keying (DP-QPSK) and channel arrangement. The DP-QPSK was designed and implemented using 100 GHz for both Equal Space Channel Allocation (ESCA) and Unequal Space Channel Allocation (USCA). It was analysed on a 64-channel DWDM system operating at 100 Gbps capacity over a transmission distance of 1000 km. The DP-QPSK-based ESCA and USCA techniques were simulated using OptiSystem 17 simulation software and Python for data analysis. The performance of the system was evaluated using bit error rate (BER), optical signal-to-noise ratio (OSNR) and Electrical Constellation Diagram. Validation was done with other advanced modulation formats: Quadrature Phase Shift Keying (QPSK) and Differential-Quadrature Phase Shift Keying (D-QPSK). The DWDM system using DP-QPSK gave better performance due to the highest OSNR and BER values obtained when compared with ESCA-based QPSK and D-QPSK techniques. The technique can be used to suppress the effect of FWM in optical communication systems.

Enhancing YOLO for occluded vehicle detection with grouped orthogonal attention and dense object repulsion

In real-life complex traffic environments, vehicles are often occluded by extraneous background objects and other vehicles, leading to severe degradation of object detector performance. To address this issue, we propose a method named YOLO-OVD (YOLO for occluded vehicle detection) and a dataset for effectively handling vehicle occlusion in various scenarios. To highlight the model attention in unobstructed region of vehicles, we design a novel grouped orthogonal attention (GOA) module to achieve maximum information extraction between channels. We utilize grouping and channel shuffling to address the initialization and computational issues of original orthogonal filters, followed by spatial attention for enhancing spatial features in vehicle-visible regions. We introduce a CIoU-based repulsion term into the loss function to augment the network’s localization accuracy in scenarios involving densely packed vehicles. Moreover, we explore the effect of the knowledge-based Laplacian Pyramid on the OVD performance, which contributes to fast convergence in training and ensures more detailed and comprehensive feature retention. We conduct extensive experiments on the established occluded vehicle detection dataset, which demonstrates that the proposed YOLO-OVD model significantly outperforms 14 representative object detectors. Notably, it achieves improvements of 4.7% in Precision, 3.6% in AP@0.5, and 1.9% in AP@0.5:0.95 compared to the YOLOv5 baseline.

Long-range attention classification for substation point cloud

Point cloud classification for substation is challenging due to the occluding layout as well as the complexity of methods. Most previous methods trade inference speed for precision by building complex extractors. To alleviate the paradox of performance and complexity trade-off, this paper proposes the lightweight long-range attention classification for substation point cloud, including shuffle channel attention(SCA) and dilated channel attention(DCA). First, SCA captures local cross-channel interaction via 1D convolution and global interaction via channel shuffling, shown promising performance. Second, to further reduce the amount of computation involved in shuffling, we propose a more elegant method DCA by resizing the 1D vector after pooling into 2D feature map. Note that proposed DCA are implemented by just 2D convolution, determining the coverage of cross-channel interaction, with a significantly higher inference speed. Furthermore, we develop a heuristic algorithm to adaptively determine parameters like kernel size, shuffle group and size of 2D feature map. Besides ModelNet40 and ScanObjectNN(PB_T50_RS), this paper also selects ten of main objects in substation for training and testing. Finally, experimental results suggest that proposed methods bring notable performance gain, nearly 1%. DCA reaches 93.801% and just increases 0.0003M parameters but 15.5% faster than SCA. We also make an attempt to simplify proposed methods via reducing the coding dimensions and coding blocks. And simplified DCA achieves 93.4% performance with almost quarter parameters and 112% faster, guaranteeing both efficiency and effectiveness. A simplified version is used to achieve the highest accuracy of 85.115% and 83.304% on ScanObjectNN, with the highest improvements of 1.735% and 2.391%. This paper also conducts robustness test where different proportions of points are missing. Results show that when 93.75% missing, the accuracy only decreases by 4.093%, and can still reach 89%, which is significantly ahead of other methods.

Shuffle-fusion pyramid network for bearing fault diagnosis under noisy environments

Abstract Recent advancements in deep learning have propelled the exploration of big data-driven fault diagnosis techniques. Nevertheless, traditional models often suffer from prohibitive computational demands, rendering them impractical for on-site deployment in rolling bearing fault diagnosis. To address this challenge, this paper introduces a novel lightweight fault diagnosis model with the pyramid architecture, named Shuffle-Fusion Pyramid Network (Shuffle-FPN). The model heralds several innovations: (1) A pyramid structure is designed to amalgamate fault signals across various scales, enlarging the network’s breadth and curtailing its depth. (2) Depth wise separable convolutions are adopted to streamline network parameters, thus achieving a lightweight model, and channel shuffling to ensure thorough information fusion across convolutional channels. (3) A global representation module is employed to offset the loss of global context, which accompanies increased convolutional depth. Collectively, these innovations enable Shuffle-FPN to extract nuanced fault features amidst noise and to operate on devices with limited memory, ensuring real-time fault diagnosis even in complex environments. Rigorous experiments on public datasets from Paderborn University (PU), supplemented by validations with our research group’s experiment data, reveal that Shuffle-FPN excels in fault identification under severe noise conditions and reduces memory footprint, which solidifies the superiority of Shuffle-FPN for practical fault diagnosis.

Lightweight safflower cluster detection based on YOLOv5

The effective detection of safflower in the field is crucial for implementing automated visual navigation and harvesting systems. Due to the small physical size of safflower clusters, their dense spatial distribution, and the complexity of field scenes, current target detection technologies face several challenges in safflower detection, such as insufficient accuracy and high computational demands. Therefore, this paper introduces an improved safflower target detection model based on YOLOv5, termed Safflower-YOLO (SF-YOLO). This model employs Ghost_conv to replace traditional convolution blocks in the backbone network, significantly enhancing computational efficiency. Furthermore, the CBAM attention mechanism is integrated into the backbone network, and a combined LCIOU+NWD\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$L_{CIOU + NWD}$$\\end{document} loss function is introduced to allow for more precise feature extraction, enhanced adaptive fusion capabilities, and accelerated loss convergence. Anchor boxes, updated through K-means clustering, are used to replace the original anchors, enabling the model to better adapt to the multi-scale information of safflowers in the field. Data augmentation techniques such as Gaussian blur, noise addition, sharpening, and channel shuffling are applied to the dataset to maintain robustness against variations in lighting, noise, and visual angles. Experimental results demonstrate that SF-YOLO surpasses the original YOLOv5s model, with reductions in GFlops and Params from 15.8 to 13.2 G and 7.013 to 5.34 M, respectively, representing decreases of 16.6% and 23.9%. Concurrently, SF-YOLO’s mAP0.5 increases by 1.3%, reaching 95.3%. This work enhances the accuracy of safflower detection in complex agricultural environments, providing a reference for subsequent autonomous visual navigation and automated non-destructive harvesting technologies in safflower operations.

YOLO-ABD: A Multi-Scale Detection Model for Pedestrian Anomaly Behavior Detection

Public safety and intelligent surveillance systems rely on anomaly detection for effective monitoring. In real-world pedestrian detection scenarios, Pedestrians often exhibit various symmetrical features such as body contours, facial features, posture, and clothing. However, the accuracy of pedestrian anomaly detection is affected by factors such as complex backgrounds, pedestrian obstruction, and small target sizes. To address these issues, this study introduced YOLO-ABD, a lightweight method for anomaly behavior detection that integrated small object detection and channel shuffling. This approach enhanced the YOLOv8n baseline model by integrating a small-object detection mechanism at the head and employing the symmetric GSConv convolutional module in the backbone network to improve perceptual capabilities. Furthermore, it incorporated the SimAM attention mechanism to mitigate complex background interference and thus enhance target detection performance. Evaluation on the IITB-Corridor dataset showed mAP50 and mAP50-95 scores of 89.3% and 60.6%, respectively. Generalization testing on the street-view-gdogo dataset further underscored the superiority of YOLO-ABD over advanced detection algorithms, demonstrating its effectiveness and generalization capabilities. With relatively fewer parameters, YOLO-ABD provided an excellent lightweight solution for pedestrian anomaly detection.

Asymmetric convolutional modulation network for efficient image super-resolution

Thanks to the effectiveness of large kernel convolutions in acquiring large receptive fields, lightweight single image super-resolution (SISR) approaches based on large kernel designs have achieved significant performance. However, these state-of-the-art methods still require high computational costs, and their model efficiency remains to be improved. To address these challenges, we propose an asymmetric convolutional modulation network (ACMN) for efficient image super-resolution. In detail, we design an asymmetric convolutional modulation unit (ACMU) that combines large kernel design and modulation mechanism for adaptively selecting representative features from a large receptive field. Furthermore, we adopt depth-wise asymmetric convolutions to further reduce computational burden. To supplement the local context information and promote inter-channel interaction, we further develop a channel shuffle feedforward network (CSFN) to extract local feature information and facilitate inter-channel information interaction. Experimental results demonstrate that our ACMN outperforms other state-of-the-art efficient SR methods with fewer parameters and Multi-Adds.

DDANF: Deep denoising autoencoder normalizing flow for unsupervised multivariate time series anomaly detection

In recent years, the proliferation of IoT technologies and the widespread adoption of wireless sensors across various critical infrastructures such as power plants, service monitoring systems, space and earth exploration missions, and water treatment facilities have resulted in the generation of vast quantities of multivariate time series data. Within this context, unsupervised anomaly detection has emerged as a pivotal yet challenging problem in time series research, necessitating machine learning models capable of identifying rare anomalies amidst massive datasets. Traditionally, unsupervised methods have approached this issue by learning representations of primary patterns within sequences and detecting deviations through reconstruction errors. However, the effectiveness of this approach is often limited due to the intricate dynamics and diverse patterns inherent in these dynamic systems. Moreover, many existing unsupervised anomaly detection techniques fail to fully exploit inter-feature relationships within multivariate time series data, thereby overlooking a crucial criterion for accurate detection. To address these shortcomings, this paper introduces a novel unsupervised method for multivariate time series anomaly detection based on normalized flows and autoencoders. Central to our approach is the incorporation of a channel shuffling mechanism during training, enhancing the model’s capacity to discern inter-channel patterns and anomalies. Concurrently, the application of normalized flows within the autoencoder framework serves to constrain the latent space, effectively isolating anomalies and improving detection accuracy. Experimental validation conducted on two large-scale public datasets demonstrates the efficacy of the proposed method compared to established benchmarks, highlighting its superior performance.

PAS-YOLO: improved YOLOv8 integrating parameter-free attention and channel shuffle for object detection of power grid equipment

PAS-YOLO: improved YOLOv8 integrating parameter-free attention and channel shuffle for object detection of power grid equipment

YOLO-Granada: a lightweight attentioned Yolo for pomegranates fruit detection

Pomegranate is an important fruit crop that is usually managed manually through experience. Intelligent management systems for pomegranate orchards can improve yields and address labor shortages. Fast and accurate detection of pomegranates is one of the key technologies of this management system, crucial for yield and scientific management. Currently, most solutions use deep learning to achieve pomegranate detection, but deep learning is not effective in detecting small targets and large parameters, and the computation speed is slow; therefore, there is room for improving the pomegranate detection task. Based on the improved You Only Look Once version 5 (YOLOv5) algorithm, a lightweight pomegranate growth period detection algorithm YOLO-Granada is proposed. A lightweight ShuffleNetv2 network is used as the backbone to extract pomegranate features. Using grouped convolution reduces the computational effort of ordinary convolution, and using channel shuffle increases the interaction between different channels. In addition, the attention mechanism can help the neural network suppress less significant features in the channels or space, and the Convolutional Block Attention Module attention mechanism can improve the effect of attention and optimize the object detection accuracy by using the contribution factor of weights. The average accuracy of the improved network reaches 0.922. It is only less than 1% lower than the original YOLOv5s model (0.929) but brings a speed increase and a compression of the model size. and the detection speed is 17.3% faster than the original network. The parameters, floating-point operations, and model size of this network are compressed to 54.7%, 51.3%, and 56.3% of the original network, respectively. In addition, the algorithm detects 8.66 images per second, achieving real-time results. In this study, the Nihui convolutional neural network framework was further utilized to develop an Android-based application for real-time pomegranate detection. The method provides a more accurate and lightweight solution for intelligent management devices in pomegranate orchards, which can provide a reference for the design of neural networks in agricultural applications.

Fast identification of tomatoes in natural environments by improved YOLOv5s

Real time recognition and detection of tomato fruit maturity is a key function of tomato picking robots. Existing recognition and detection algorithms have slow speed and low recognition accuracy for small tomatoes. Here, a tomato fruit maturity detection model YOLOv5s3 based on improved YOLOv5s was proposed and its accuracy was verified through comparative experiments. On the basis of YOLOv5s, an SC module was proposed based on channel shuffle packet convolution. Then, A C3S module is constructed, which replaced the original C3 module with this C3S module to reduce the number of parameters while maintaining the feature expression ability of the original network. And a 3-feature fusion FF module was put forward, which accepted inputs from three feature layers. The FF module fused two feature maps from the backbone network. The C2 layer of the backbone was integrated, and the large target detection head was removed to use dual head detection to enhance the detection ability of small targets. The experimental results showed that the improved model has a detection accuracy of 94.8%, a recall rate of 96%, a parameter quantity of 3.02M, and an average accuracy (mAP0.5) of 93.3% for an intersection over union (IoU) of 0.5. The detection speed reaches 9.4ms. It can quickly and accurately identify the maturity of tomato fruits, and the detection speed is 22.95%, 33.33%, 48.91%, 68.35%, 15%, and 25.98% higher than the original YOLOv5s, YOLOv5m, YOLOv5l, YOLOv5x, YOLOv5n, and YOLOv4, respectively. The real-time testing visualization results of different models indicated that the improved model can effectively improve detection speed and solve the problem of low recognition rate for small tomatoes, which can provide reference for the development of picking robots.

Different life cycles of rice pests’ images recognition based on adaptive lightweight DC-ghost module

Timely protection and treatment of rice crops require accurate identification of pests. For this purpose, this paper introduces a lightweight network called DC-GhostNet (DO-Channel Shuffle GhostNet). This network combines DO convolution, feature extraction and fusion, and IECA lightweight attention mechanism with adaptive k-values to identify rice pests effectively. In the Ghost Module, the DO convolution is used to replace the DW convolution in the Ghost convolution to enhance the feature extraction more efficiently, without increasing the model parameters and computational effort. To conquer the challenge of rice pest image feature extraction in different complex scenes, the channel shuffle operation is added at the end of the Ghost Module to extract deep global features and optimized them. The model is then subjected to a transfer learning approach to highly boost the network’s performance. Experiments show that our method’s accuracy, precision, recall, and F1 scores on the test dataset of 15 types of rice pests are, 89.15 %, 87.13 %, 85.77 %, and 86.45 %, respectively, which is 7.44 percentage points, 6.42 percentage points, 11.96 percentage points, and 9.35 percentage points higher than the baseline GhostNet. The DC-GhostNet model contains 4.35 M parameters and 141.99 M floating point operations FLOPs, which are 16 % and 9 % less than GhostNet, respectively. The performance of the proposed model under different life cycles of ovum, larvae, and adults is also optimal compared to other networks, and the final inference time of the model for a single image is under 13.1 ms, demonstrating both its speedy detection and lightweight deployment capabilities.

SOCR‐YOLO: Small Objects Detection Algorithm in Medical Images

ABSTRACTIn the field of medical image analysis, object detection plays a crucial role by providing interpretable diagnostic information to healthcare professionals. Although current object detection models have achieved remarkable success in conventional images, their performance in detecting abnormalities in medical images has not been as satisfactory. This is primarily due to the complexity of anatomical structures in medical images, and the fact that some lesions may have subtle features, particularly in the case of early‐stage, small‐scale abnormalities. To address this challenge, we introduce SOCR‐YOLO, a novel lesion detection model with online convolutional reparameterization based on channel shuffling. First, it employs the SOCR (Shuffled Channel with Online Convolutional Re‐parameterization) module to establish a connection between feature concatenation and computational efficiency, aiming to extract more comprehensive information while reducing time consumption. Second, it incorporates the Bi‐FPN structure to achieve multiscale feature fusion. Lastly, the loss function has been optimized to improve the model training process. We evaluated two datasets, chest x‐ray (Vindr‐CXR) and brain tumor (Br35H), provided by the Kaggle competition. Experimental results show that the proposed method has outperformed several state‐of‐the‐art models, including YOLOv8, YOLO‐NAS, and RT‐DETR, in both speed and accuracy. Notably, in the context of chest x‐ray anomaly detection, SOCR‐YOLO exhibits a 1.8% enhancement in accuracy over YOLOv8 while simultaneously reducing floating‐point operations by 26.3%. Additionally, a similar 1.8% improvement in accuracy is observed in the detection of brain tumors. The results indicate the superior ability of our model to detect multiscale variations and small lesions.