Mean Average Precision Research Articles

Unraveling Indonesian heritage through pattern recognition using YOLOv5

This research focuses on three iconic Indonesian batik patterns-Kawung, Mega Mendung, and Parang-due to their cultural significance and recognition. Kawung symbolizes harmony, Mega Mendung represents power, and Parang signifies protection and spiritual power. Using the YOLOv5 deep learning model, the study aimed to accurately identify these patterns. Results showed mean average precision (mAP) scores of 77% for Kawung, 80% for Parang, and an impressive 99% for Mega Mendung. The highest precision results were 91% for Kawung, 88% for Parang, and 77% for Mega Mendung. These findings highlight the potential of pattern recognition in preserving cultural heritage. Understanding these designs contributes to the appreciation of Indonesia s culture. The research suggests applications in cultural studies, digital archiving, and the textile industry, ensuring the legacy of these patterns endures.

The Development Mask R-CNN Model for Identification of Melon Plant Leaves and Branches

The quality of melons can be enhanced and optimized by pruning melon plants. Pruning is a removal process carried out on specific parts of the plant. Currently, melon plants are still pruned manually by farmers, but there are many drawbacks to this method. In this research, pruning is conducted on the branches and leaves of melon plants. Pruning can be facilitated with the assistance of a robot capable of recognizing leaves and branches. In this study, the method used to detect branches and leaves is the Mask Region-based Convolutional Neural Network (Mask R-CNN). Hyperparameter tuning technique is employed to obtain the best parameter values, including learning rate, weight decay, and learning momentum. Two scenarios are considered in this research, one with 10 epochs and the other with 30 epochs. The obtained Average Precision (AP) values at 10 epochs are 32.2% for leaf objects and 0% for branches. At 30 epochs, the AP values are 56.8% for leaf objects and 4.1% for branches. The mean Average Precision (mAP) is 16.1% for 10 epochs and 28.4% for 30 epochs.

A Study of Word Bigrams for Pseudo-relevance Feedback in Information Retrieval

Traditional information retrieval models mostly adopt a term independence assumption and are based on single terms or unigrams. Past efforts have attempted to go beyond this assumption, such as by using contiguous terms (i.e. word n-grams) or terms appearing in proximity. One such approach employs pseudo-relevance feedback (PRF) in an extended BM25 model, with an expanded query containing bigrams and proximity word pairs besides unigrams. However, the benefit of this approach over the traditional unigram PRF remains inconclusive. We speculate the uncertain effectiveness of bigram PRF in this past work is due to: (1) The new bigrams obtained from the expanded query may be formed by pairing unigrams drawn from different documents. These are potentially noise instead of relevant concepts; (2) The collection statistics of n-grams needed to calculate the document ranking functions, such as their document frequency, is not available in retrieval. Only estimates of these quantities are used instead. We suggest that these issues may be overcome by extracting word n-grams as single units in query expansion, and employing a document index that contains both unigrams and word n-grams. We demonstrate the approach for the case of bigram PRF in an extended BM25. Retrieval experiments are conducted on a range of standard test collections. For the majority of tested collections, the difference between values of the evaluation metrics (Mean Average Precision and the precision-oriented NDCG@20) obtained by our bigram PRF and the unigram PRF baseline is not statistically significant. Thus, our bigram PRF fails to improve over unigram PRF robustly across collections. An analysis of our results reveals ‘query drift’ due to bigram query expansion terms that represent too-broad topics as a cause for the failure of our approach.

Facial Expression Recognition-You Only Look Once-Neighborhood Coordinate Attention Mamba: Facial Expression Detection and Classification Based on Neighbor and Coordinates Attention Mechanism

In studying the joint object detection and classification problem for facial expression recognition (FER) deploying the YOLOX framework, we introduce a novel feature extractor, called neighborhood coordinate attention Mamba (NCAMamba) to substitute for the original feature extractor in the Feature Pyramid Network (FPN). NCAMamba combines the background information reduction capabilities of Mamba, the local neighborhood relationship understanding of neighborhood attention, and the directional relationship understanding of coordinate attention. The resulting FER-YOLO-NCAMamba model, when applied to two unaligned FER benchmark datasets, RAF-DB and SFEW, obtains significantly improved mean average precision (mAP) scores when compared with those obtained by other state-of-the-art methods. Moreover, in ablation studies, it is found that the NCA module is relatively more important than the Visual State Space (VSS), a version of using Mamba for image processing, and in visualization studies using the grad-CAM method, it reveals that regions around the nose tip are critical to recognizing the expression; if it is too large, it may lead to erroneous prediction, while a small focused region would lead to correct recognition; this may explain why FER of unaligned faces is such a challenging problem.

Breast Lesion Detection for Ultrasound Images Using MaskFormer

This study evaluates the performance of the MaskFormer model for segmenting and classifying breast lesions using ultrasound images, addressing ultrasound’s limitations. Ultrasound used for breast cancer detection faces challenges like low image contrast and difficulty in the detection of small or multiple lesions, further complicated by variability based on operator skill. Initial experiments with U-Net and other CNN-based models revealed constraints, such as early plateauing in model loss, which indicated suboptimal learning and performance. In contrast, MaskFormer demonstrated continuous improvement, achieving higher precision in breast lesion segmentation and significantly reducing both false positives and false negatives. Comparative analysis showed MaskFormer’s superior performance, with the highest precision and recall rates for malignant lesions and an overall mean average precision (mAP) of 0.943. The model’s ability to detect a diverse range of breast lesions, including those potentially missed by the human eye, especially by less experienced practitioners, underscores its potential. These findings suggest that integrating AI models like MaskFormer could greatly enhance ultrasound performance for breast cancer detection, providing reliable, operator-independent image analysis and potentially improving patient outcomes on a global scale.

Research and Experiment on Miss-Seeding Detection of Potato Planter Based on Improved YOLOv5s

In order to improve the performance of potato planter, reduce miss-seeding rates, enhance the overall quality of the seeding operation, and ultimately increase the yield of the potato, it is necessary to implement effective technical means to monitor and identify the miss-seeding issues during the seeding process. The existing miss-seeding detection technologies commonly use sensors to monitor, but such technologies are easily affected by factors like heavy dust and strong vibrations, resulting in poor interference resistance and adaptability. Therefore, this study aims to explore and apply deep learning algorithms to achieve real-time monitoring of the miss-seeding phenomenon in potato planter during the planting process. Considering both the lightweight of the miss-seeding detection model and its practical deployment, this study selects and adapts the YOLOv5s algorithm to achieve this goal. Firstly, the attention mechanism is integrated into the backbone network to suppress background interference and improve detection accuracy. Secondly, the non-maximum suppression algorithm is improved by replacing the original IoU-NMS with the Soft-NMS algorithm to enhance the bounding box regression rate and reduce missed detections of potato seeds due to background overlap or occlusion. Experimental results show that the accuracy of the improved algorithm in detecting miss-seeding increased from 96.02% to 98.30%, the recall rate increased from 96.31% to 99.40%, and the mean average precision (mAP) improved from 99.12% to 99.40%. The improved model reduces missed and false detections, provides more precise target localization, and is suitable for miss-seeding detection in natural environments for potato planter, providing technical and theoretical support for subsequent intelligent reseeding in potato planter.

BHI-YOLO: A Lightweight Instance Segmentation Model for Strawberry Diseases

In complex environments, strawberry disease segmentation models face challenges, such as segmentation difficulties, excessive parameters, and high computational loads, making it difficult for these models to run effectively on devices with limited computational resources. To address the need for efficient running on low-power devices while ensuring effective disease segmentation in complex scenarios, this paper proposes BHI-YOLO, a lightweight instance segmentation model based on YOLOv8n-seg. First, the Universal Inverted Bottleneck (UIB) module is integrated into the backbone network and merged with the C2f module to create the C2f_UIB module; this approach reduces the parameter count while expanding the receptive field. Second, the HS-FPN is introduced to further reduce the parameter count and enhance the model’s ability to fuse features across different levels. Finally, by integrating the Inverted Residual Mobile Block (iRMB) with EMA to design the iRMA, the model is capable of efficiently combining global information to enhance local information. The experimental results demonstrate that the enhanced instance segmentation model for strawberry diseases achieved a mean average precision (mAP@50) of 93%. Compared to YOLOv8, which saw a 2.3% increase in mask mAP, the improved model reduced parameters by 47%, GFLOPs by 20%, and model size by 44.1%, achieving a relatively excellent lightweight effect. This study combines lightweight architecture with enhanced feature fusion, making the model more suitable for deployment on mobile devices, and provides a reference guide for strawberry disease segmentation applications in agricultural environments.

Improved Real-Time Detection Transformer-Based Rail Fastener Defect Detection Algorithm

To address the issues of the Real-Time DEtection TRansformer (RT-DETR) object detection model, including poor defect feature extraction in the task of rail fastener defect detection, inefficient use of computational resources, and suboptimal channel attention in the self-attention mechanism, the following improvements were made. Firstly, a Super-Resolution Convolutional Module (SRConv) was designed as a separate component and integrated into the Backbone network, which enhances the image details and clarity while preserving the original image structure and semantic content. This integration improves the model’s ability to extract defect features. Secondly, a channel attention mechanism was integrated into the self-attention module of RT-DETR to enhance the focus on feature map channels, addressing the problem of sparse attention maps caused by the lack of channel attention while saving computational resources. Finally, the experimental results show that compared to the original model, the improved RT-DETR-based rail fastener defect detection algorithm, with an additional 0.4 MB of parameters, achieved a higher accuracy, with a 2.8 percentage point increase in the Mean Average Precision (mAP) across IoU thresholds from 0.5 to 0.9 and a 1.7 percentage point increase in the Average Recall (AR) across the same thresholds.

EMB-YOLO: A Lightweight Object Detection Algorithm for Isolation Switch State Detection

In power inspection, it is crucial to accurately and regularly monitor the status of isolation switches to ensure the stable operation of power systems. However, current methods for detecting the open and closed states of isolation switches based on image recognition still suffer from low accuracy and high edge deployment costs. In this paper, we propose a lightweight object detection model, EMB-YOLO, to address this challenge. Firstly, we propose an efficient mobile inverted bottleneck convolution (EMBC) module for the backbone network. This module is designed with a lightweight structure, aimed at reducing the computational complexity and parameter count, thereby optimizing the model’s computational efficiency. Furthermore, an ELA attention mechanism is used in the EMBC module to enhance the extraction of horizontal and vertical isolation switch features in complex environments. Finally, we proposed an efficient-RepGDFPN fusion network. This network integrates feature maps from different levels to detect isolation switches at multiple scales in monitoring scenarios. An isolation switch dataset was self-built to evaluate the performance of the proposed EMB-YOLO. The experimental results demonstrated that the proposed method achieved superior detection performance on our self-built dataset, with a mean average precision (mAP) of 87.2%, while maintaining a computational cost of only 6.5×109 FLOPs and a parameter size of just 2.8×106 bytes.

Colon polyp detection based on multi-scale and multi-level feature fusion and lightweight convolutional neural network

Early diagnosis and treatment of colorectal polyps are crucial for preventing colorectal cancer. This paper proposes a lightweight convolutional neural network for the automatic detection and auxiliary diagnosis of colorectal polyps. Initially, a 53-layer convolutional backbone network is used, incorporating a spatial pyramid pooling module to achieve feature extraction with different receptive field sizes. Subsequently, a feature pyramid network is employed to perform cross-scale fusion of feature maps from the backbone network. A spatial attention module is utilized to enhance the perception of polyp image boundaries and details. Further, a positional pattern attention module is used to automatically mine and integrate key features across different levels of feature maps, achieving rapid, efficient, and accurate automatic detection of colorectal polyps. The proposed model is evaluated on a clinical dataset, achieving an accuracy of 0.9982, recall of 0.9988, F1 score of 0.9984, and mean average precision (mAP) of 0.9953 at an intersection over union (IOU) threshold of 0.5, with a frame rate of 74 frames per second and a parameter count of 9.08 M. Compared to existing mainstream methods, the proposed method is lightweight, has low operating configuration requirements, high detection speed, and high accuracy, making it a feasible technical method and important tool for the early detection and diagnosis of colorectal cancer.

EVIT-YOLOv8: Construction and research on African Swine Fever facial expression recognition

The threats of infectious diseases such as African Swine Fever, Swine Erysipelas, and Blue Ear Disease to the pig farming industry have been increasing year by year. Producers often face difficulties in diagnosis, leading to the misuse of measures and the spread of epidemics. Addressing this, the proposed EVIT-YOLOv8 model integrates the EViT module to surpass ViT limitations and incorporates the CBAM module for enhanced image feature representation. Employing the GIOU loss function ensures better precision in capturing facial expression features, yielding an impressive Mean Average Precision (mAP) of 86.6% in differentiation tasks. Specifically, in African Swine Fever facial expression recognition, the model achieves a remarkable Precision of 85.2%, outperforming YOLOv5, YOLOv7, and YOLOv8 models by 6%, 23.5%, and 7.3%, respectively. This provides pig producers with a precise diagnostic tool, mitigating the risk of epidemic spread due to misdiagnosis and facilitating effective prevention and control of infectious diseases.

Weak saliency ensemble network for person Re-identification using infrared light images

In recent years, person re-identification (re-id) has primarily been studied using visible light (VL) images. However, the challenges of employing VL images in nighttime environments have prompted research into using infrared light (IR) images. Yet, the utilization of both VL and IR images in person re-id has resulted in increased computational cost and processing time in multi-modality systems, leading to studies focusing solely on IR images. Nevertheless, IR images, lacking color and texture information, generally yield lower recognition performance in existing person re-id studies. In addition, previous studies have shown that person re-id performance suffers in the presence of complex background noise. To tackle these challenges, this study proposes a new weak saliency ensemble network (WSE-Net) for person re-id using IR images. WSE-Net incorporates a channel reduction of feature (CRF) method to reduce computational cost in the ensemble network, a technique for converting input images into group of patch images and feeding them into the ensemble model to enhance the reduced feature information, and a grouped convolution ensemble network (GCE-Net) that enables the fusion of features extracted from original and attention-guided ensemble models.The performance of person re-id using WSE-Net was evaluated on the Dongguk body-based person recognition database version 1 (DBPerson-Recog-DB1) and the Sun Yat-sen university multiple modality re-identification version 1 (SYSU-MM01). Experimental results demonstrated that on DBPerson-Recog-DB1, WSE-Net achieved 93.65% in rank 1, 95.28% in mean average precision (mAP), and 93.52% in the harmonic mean of precision and recall. Additionally, on SYSU-MM01, WSE-Net achieved 86.85% in rank 1, 44.58% in mAP, and 40.06% in the harmonic mean of precision and recall. Furthermore, the accuracy of WSE-Net on both datasets surpassed that of state-of-the-art methods.

CVW-Etr: A High-Precision Method for Estimating the Severity Level of Cotton Verticillium Wilt Disease

Cotton verticillium wilt significantly impacts both cotton quality and yield. Selecting disease-resistant varieties and using their resistance genes in breeding is an effective and economical control measure. Accurate severity estimation of this disease is crucial for breeding resistant cotton varieties. However, current methods fall short, slowing the breeding process. To address these challenges, this paper introduces CVW-Etr, a high-precision method for estimating the severity of cotton verticillium wilt. CVW-Etr classifies severity into six levels (L0 to L5) based on the proportion of segmented diseased leaves to lesions. Upon integrating YOLOv8-Seg with MobileSAM, CVW-Etr demonstrates excellent performance and efficiency with limited samples in complex field conditions. It incorporates the RFCBAMConv, C2f-RFCBAMConv, AWDownSample-Lite, and GSegment modules to handle blurry transitions between healthy and diseased regions and variations in angle and distance during image collection, and to optimize the model’s parameter size and computational complexity. Our experimental results show that CVW-Etr effectively segments diseased leaves and lesions, achieving a mean average precision (mAP) of 92.90% and an average severity estimation accuracy of 92.92% with only 2.6M parameters and 10.1G FLOPS. Through experiments, CVW-Etr proves robust in estimating cotton verticillium wilt severity, offering valuable insights for disease-resistant cotton breeding applications.

Automated classification in turtles genus Malayemys using ensemble multiview image based on improved YOLOv8 with CNN

In Thailand, two snail-eating turtle species in the genus Malayemes (M. subtrijuga and M. macrocephala) are protected animals in which smuggling and trading are illegal. Recently, a new species M. khoratensis has been reported and it has not yet been considered as protected animal species. To enforce the law, species identification of Malayemes is crucial. However, it is quite challenging and requires expertise. Therefore, a simple tool, such as image analysis, to differentiate these three snail-eating species would be highly useful. This study proposes a novel ensemble multiview image processing approach for the automated classification of three turtle species in the genus Malayemys. The original YOLOv8 architecture was improved by utilizing a convolutional neural network (CNN) to overcome the limitations of traditional identification methods. This model captures unique morphological features by analyzing Malayemys species images from various angles, addressing challenges such as occlusion and appearance variations. The ensemble multiview strategy significantly increases the YOLOv8 classification accuracy using a comprehensive dataset, achieving an average mean average precision (mAP) of 98% for the genus Malayemys compared with the nonensemble multiview and single-view strategies. The species identification accuracy of the proposed models was validated by comparing genetic methods using mitochondrial DNA with morphological characteristics. Even though the morphological characteristics of these three species are ambiguous, the mitochondrial DNA sequences are quite distinct. Therefore, this alternative tool should be used to increase confidence in field identification. In summary, the contribution of this study not only marks a significant advancement in computational biology but also supports wildlife and turtle conservation efforts by enabling rapid, accurate species identification.

Simultaneous segmentation and classification of 99mTc-DMSA renal scintigraphic images with a deep learning approach.

99mTc-DMSA scan plays an important role in assessing functional abnormalities in kidneys. As a promising network for deep learning (DL), Mask R-CNN has the capability of simultaneously segmenting and classifying objects in images. In this study, we tested the feasibility and accuracy of Mask R-CNN in diagnosing acute pyelonephritis (APN) and segmenting kidneys in 99mTc-DMSA scintigraphic images. Two hundred and sixty patients with suspected APN were recruited for DMSA scan, of which 358 kidneys were diagnosed as APN. Of the recruited patients, 210 were randomly selected for training and validating Mask R-CNN, and the other 50 patients' images were used for model testing. Accuracy of the results was assessed by comparing against references from human experts. In the validation phase, the trained model provided segmentation masks with intersection over union (mask IoU, for segmentation accuracy) of 86.6%, and classifications with mean average precision at the bounding box IoU ≥ 50% (mAP50, for classification accuracy) of 86.2%. In testing with the 50 independent patients, mask IoU of the model's segmentation was 90.3%±2.2%, and in classifying the kidneys for APN, the trained model showed accuracy of 89.0%, sensitivity of 84.8% and specificity of 97.0%. In identifying patients with any APN kidney, 3 out of 50 patients were mis-diagnosed, achieving accuracy of 94.0%. Mask R-CNN, designed to perform both segmentation and classification for images, showed much promise in analyzing 99mTc-DMSA images for both accurate diagnosis of APN and kidney segmentation.

Deep Learning Approach for Arm Fracture Detection Based on an Improved YOLOv8 Algorithm

Artificial intelligence (AI)-assisted computer vision is an evolving field in medical imaging. However, accuracy and precision suffer when using the existing AI models for small, easy-to-miss objects such as bone fractures, which affects the models’ applicability and effectiveness in a clinical setting. The proposed integration of the Hybrid-Attention (HA) mechanism into the YOLOv8 architecture offers a robust solution to improve accuracy, reliability, and speed in medical imaging applications. Experimental results demonstrate that our HA-modified YOLOv8 models achieve a 20% higher Mean Average Precision (mAP 50) and improved processing speed in arm fracture detection.

An Improved YOLOv9 and Its Applications for Detecting Flexible Circuit Boards Connectors

Flexible circuit boards are a cornerstone of the modern electronics industry. In automatic defect detection, FPC connectors present challenges such as minimal differences between oxidation defects and the background, easy degradation of Intersection over Union (IoU) scores, and significant variations in the shapes of black defect boundaries. Consequently, existing algorithms perform poorly in this task. We improve model YOLOv9 by introducing Multi-scale Dilated Attention (MSDA) on the output side to enhance the ability to capture features, and Deformable Large Kernel Attention (DLKA) on the other side of the output header to improve the ability to adapt to complex defect boundaries. Our use of IoU loss completely eliminates the risk of IoU degradation or gradient vanishing. Furthermore, we reduce computational overhead with the implementation of Faster Block. Following these improvements, the mean Average Precision (mAP) at 75% IoU (mAP75) for oxidized defects increased by 7.5% relative to the base model. Similarly, the mAP at 50% IoU (mAP50) for black defects increased by 5.7%, validating the relevance and efficacy of our proposed improvements. Overall, the average mAP50, mAP75, and mAP50:95 for all defects improved by 3.8%, 2.0%, and 2.3%, respectively. The performance gain achieved by our enhanced model significantly exceeds the improvement of YOLOv9 relative to YOLOv8.

Sparse cross-transformer network for surface defect detection.

Quality control processes with automation ensure that customers receive defect-free products that meet their needs. However, the performance of real-world surface defect detection is often severely hindered by the scarcity of data. Recently, few-shot learning has been widely proposed as a solution to the data sufficiency problem by leveraging a limited number of base class samples. However, achieving discriminative and generalization capabilities with few samples remains a challenging task in various surface defect detection scenarios. In this paper, we propose a sparse cross-transformer network (SCTN) for surface defect detection. Specifically, we introduce a residual layer module to enhance the network's ability to retain crucial information. Next, we propose a sparse layer module within the cross-transformer to increase computational efficiency. Finally, we incorporate a squeeze-and-excitation network into the cross-transformer to enhance the attention mechanism between local patches outputted by the transformer encoder. To verify the effectiveness of our proposed method, we conducted extensive experiments on the cylinder liner defect dataset, the NEU steel surface defect dataset, and the PKU-Market-PCB dataset, achieving the best mean average precision of 62.73%, 85.29%, and 88.7%, respectively. The experimental results demonstrate that our proposed method achieves significant improvements compared to state-of-the-art algorithms. Additionally, the results indicate that SCTN enhances the network's discriminative ability and effectively improves generalization across various surface defect detection tasks.

FGYOLO: An Integrated Feature Enhancement Lightweight Unmanned Aerial Vehicle Forest Fire Detection Framework Based on YOLOv8n

To address the challenges of complex backgrounds and small, easily confused fire and smoke targets in Unmanned Aerial Vehicle (UAV)-based forest fire detection, we propose an improved forest smoke and fire detection algorithm based on YOLOv8. Considering the limited computational resources of UAVs and the lightweight property of YOLOv8n, the original model of YOLOv8n is improved, the Bottleneck module is reconstructed using Group Shuffle Convolution (GSConv), and the residual structure is improved, thereby enhancing the model’s detection capability while reducing network parameters. The GBFPN module is proposed to optimize the neck layer network structure and fusion method, enabling the more effective extraction and fusion of pyrotechnic features. Recognizing the difficulty in capturing the prominent characteristics of fire and smoke in a complex, tree-heavy environment, we implemented the BiFormer attention mechanism to boost the model’s ability to acquire multi-scale properties while retaining fine-grained features. Additionally, the Inner-MPDIoU loss function is implemented to replace the original CIoU loss function, thereby improving the model’s capacity for detecting small targets. The experimental results of the customized G-Fire dataset reveal that FGYOLO achieves a 3.3% improvement in mean Average Precision (mAP), reaching 98.8%, while reducing the number of parameters by 26.4% compared to the original YOLOv8n.

On-orbit Moving Target Detection Method Based on Dual Images from Taijing-IV 01 Satellite

Abstract. With the evolution of satellite video technology, the domain has garnered increasing attention. Concurrently, advancements in deep learning have yielded numerous outcomes in target detection. This paper introduces a novel method for detecting moving targets, offering a broader detection range compared to traditional satellite video techniques, facilitating orbital target recognition from dual panchromatic image strips. Our experimental setup on the Taijing-IV 01 satellite, launched on February 27, 2022, successfully acquired two image strips separated by one second. These strips contain speed and directional information of moving objects, extractable through the frame differencing technique. We propose combining frame differencing with lightweight deep learning for target detection, extracting regions of interest (ROIs) to focus on areas with potential moving targets. This approach reduces the workload of wholeimage target detection, decreasing data processing volume by 89%. By optimizing the YOLOv8 network and using techniques like feature map fusion of low-level and high-resolution features, we enhance sensitivity to small targets. Consequently, the model size is reduced by 79%, the mean Average Precision (mAP) increases by approximately 1.8% and 4.5%, and detection speed rises by 26%. This method introduces a new paradigm in remote sensing data services, facilitating rapid acquisition and real-time transmission of positions and image information of moving targets to the ground. This significantly reduces bandwidth requirements for transmitting remote sensing information, presenting a novel strategy for data acquisition and processing in large-scale Earth observation systems and geoscientific applications.