Box Size Research Articles

Leading hadronic contribution to the muon g − 2 from lattice QCD

We compute the leading order hadronic vacuum polarization contribution to the anomalous magnetic moment of the muon. The calculations are performed using four flavors of stout smeared staggered quarks, with quark masses at their physical values. The continuum limit is taken using six different lattice spacings ranging from 0.132 fm down to 0.064 fm. All strong isospin breaking and electromagnetic effects are accounted for to leading order. A controlled infinite volume limit is taken thanks to dedicated simulations performed in box sizes up to 11 fm. Putting all these ingredients together, we find [(gµ − 2)/2]LO−HVP = 707.5[5.5] 10−10, which has a total uncertainty of 0.8%. Compared to determinations based on the e+e− → hadrons cross section, our result significantly reduces the tension between the standard model prediction for the muon g − 2 and its experimental value.

CNN-Based Multi-Object Tracking Networks with Position Correction and IMM in Intelligent Transportation System

Abstract: In multi-object tracking of Intelligent Transportation System, there are objects of different sizes in images or videos, especially pedestrian and traffic lights with low resolution . Meantime, objects are subject to occlusion and loss in object tracking. All of the above mentioned situations may lead to unsatisfactory multi-object tracking results. Attracted by the effect of deep convolution neural networks, the paper proposes a multi-object tracking network, CNN-based Multi-Object Tracking Networks with Position Correction and IMM (CNN_PC_IMM) to solve those problems. Our proposed method consists of object detection module and object tracking module. Compared to other networks, our proposed network has several main contributions that play an essential role in achieving state-of-the-art object tracking performance. In detection phase, feature fusion technique is used. We add a scale branch to the YOLOv3 network to increase the accuracy of small object prediction, and import a residual structure to enhance gradient propagation and avoid gradient disappearance or explosion for the whole network. In addition, we determine the size of the anchor box based on the size of the object in the dataset to better detection and tracking the objects. In tracking phase, IMM is used to calculate the motion state information of the object at a certain moment. Next, the optimization algorithm is proposed to fine-tune object position when the tracking object is occluded due to dense multi-object in traffic scenes or lost due to incomplete object information. Finally, experimental results and analysis are performed on MOT16 benchmark dataset with several popular tracking algorithms used to compare the performance with the proposed algorithm in the paper. It is demonstrated that the proposed algorithm has better performance on MOPA, MOTP, ML.

Driver identification and fatigue detection algorithm based on deep learning.

In order to avoid traffic accidents caused by driver fatigue, smoking and talking on the phone, it is necessary to design an effective fatigue detection algorithm. Firstly, this paper studies the detection algorithms of driver fatigue at home and abroad, and analyzes the advantages and disadvantages of the existing algorithms. Secondly, a face recognition module is introduced to crop and align the acquired faces and input them into the Facenet network model for feature extraction, thus completing the identification of drivers. Thirdly, a new driver fatigue detection algorithm based on deep learning is designed based on Single Shot MultiBox Detector (SSD) algorithm, and the additional layer network structure of SSD is redesigned by using the idea of reverse residual. By adding the detection of drivers' smoking and making phone calls, adjusting the size and number of prior boxes of SSD algorithm, improving FPN network and SE network, the identification and verification of drivers can be realized. The experimental results showed that the number of parameters decreased from 96.62 MB to 18.24 MB. The average accuracy rate increased from 89.88% to 95.69%. The projected number of frames per second increased from 51.69 to 71.86. When the confidence threshold was set to 0.5, the recall rate of closed eyes increased from 46.69% to 65.87%, that of yawning increased from 59.72% to 82.72%, and that of smoking increased from 65.87% to 83.09%. These results show that the improved network model has better feature extraction ability for small targets.

Innovative design and application of the Re-gas Circular Express Box in a double carbon background

In recent years, with the vigorous development of e-commerce, China's express delivery industry has experienced a peak of development. However, the production of express garbage during this process has led to serious environmental pollution problems. The relevant national policies have also set higher requirements for the reduction, standardization, and recycling of express boxes. Therefore, it is urgent to reduce the size of the express box. It has been observed that many express delivery companies have joined the "green revolution" of promoting express delivery boxes. However, due to factors such as cost and safety, the utilization rate of circular express delivery boxes has not improved. Based on the current situation analysis, this study will address the existing problems and challenges of the current circulating box in practical applications. It will consider designing a green, shared, and circulating express box that can be implemented with existing technology. Therefore, this study is based on RFID technology to conduct innovative research and development, design a circular shared express box named Re-gas, and establish a closed-loop system for the supply chain of express boxes. This will promote the green and efficient transformation of the express industry. However, because the designed and manufactured express box samples have not been tested, and due to limitations imposed by application scenarios, cost control, and industry barriers, the project may have some defects and challenges.

OFCOS: An Oriented Anchor-Free Detector for Ship Detection in Remote Sensing Images

Ship detection is a significant and challenging task in remote sensing. At present, anchor-based ship detectors have achieved remarkable results, but they require introducing additional parameters and their performance is easily affected by the size of anchor boxes. In this paper, we propose an anchor-free rotated detector (OFCOS) for ship detection based on FCOS, which can be trained end-to-end. Specifically, a feature-enhanced feature pyramid network (FE-FPN) is proposed, in which the structure of the feature pyramid is optimized and an attention mechanism is introduced during fusion to enhance the significance of object features. Then, to better describe the orientation of objects, a regression branch with orientation characterization capability is constructed, and a center-to-corner bounding box prediction strategy is used to improve the accuracy of object localization. Moreover, the calculation of center-ness is optimized so that the assignment of center weights is orientation-aware and adaptive to down-weight low-quality predictions. A new remote sensing ship dataset, named RS-Ship, is constructed to further verify the effectiveness and robustness of OFCOS. Our experiments show that OFCOS achieves AP values of 91.07% and 97.05% on the publicly available dataset HRSC2016 and our self-built RS-Ship dataset, respectively, which are 13.01% and 9.84% higher than FCOS. OFCOS outperforms other mainstream detection methods in terms of both detection speed and detection accuracy.

Hydrodynamic radius of dendrimers in solvents.

The diffusion properties and hydrodynamic radius, Rh, of macromolecules are important for theoretical studies and practical application. Moreover, comparison of Rh values obtained from simulation and experimental data is used to check the correctness of simulation results. Here, we study the translation mobility of poly(butylcarbosilane) dendrimers in chloroform solution using molecular dynamics simulations and consider simulation details that may influence the accuracy of the result. Different methods to estimate Rh for a dendrimer are discussed with comparison to our experimental data. It was shown that the traditional MD simulation method for extraction of the diffusion coefficient (and calculation of Rh) of dendrimers as a rule faces difficulties and requires simulation resources several times greater than, for example, the same for a linear analogue. In the majority of MD simulation papers, the diffusion coefficient and/or Rh are calculated incorrectly. Also, we establish that correction of Rh according to the simulation box or estimation of Rh by using the gyration radius does not give values close to experimental data. To avoid the mentioned problems, we found an alternative way: to consider rotational diffusion, which gives an Rh similar to that from experiment and is practically independent of the size of the simulation box and other simulation parameters.

ANS: Assimilating Near Similarity at High Accuracy for Significant Deduction of CNN Storage and Computation

Activation data size has been roaring with the development of convolutional neural networks, which accounts for the boosting storage requirements. Our insight indicates that non-zero values dominate activations, of which the patterns demonstrate near similarity. We propose ANS method to compress activations in real time during both training and inference. High compression ratio with less accuracy loss is achieved by our optimization strategies, including determination of selection box (SB) size according to the amount of zero values of layer, learning and calibrating threshold dynamically, using the mean value of similar SB as compression value. Over 49% of compression ratio is achieved with accuracy loss of less than 0.892%, as well as reduction of multiplications by more than 60%. Comparing to three state-of-art compressed methods under five mainstream CNN models, ANS provides compression ratio improvement of 3.2x over RLC5, 1.9x over GRLC and 1.7x over ZVC. The ANS compressor and decompressor are implemented in Verilog and synthesized in 28nm node, which indicates that ANS has less cost of performance and hardware overburden. ANS modules could be seamlessly attached at the interface or deeply coupled into DNN accelerator with changed data path in the MAC array, which achieve 38% and 56% reduction in energy consumption, respectively.

A Sidelobe-Aware Small Ship Detection Network for Synthetic Aperture Radar Imagery

Ship detection from synthetic aperture radar (SAR) remote sensing images is essential for monitoring water traffic and marine safety. Numerous methods for ship detection have been developed; however, the detection of small ships presents unique challenges. SAR image characteristics, such as the sidelobe effect and blurred outline induced by the special imaging mechanism, as well as the small ship size, are the primary factors that lower the detection accuracy. This paper provides a sidelobe-aware small ship detection network for synthetic aperture radar imagery. First, considering the sidelobe effect and blurred outline, dual-pooling, i.e., average pooling and max pooling, was utilized to build a feature extraction module that lowered the effects of strong scattering points outside of the ship body and enhanced the ship body information. Second, as the bipartition process of the average pooling and maximum pooling caused some loss of original data information, different feature maps in the network were concatenated to construct a new network structure to compensate for the information lost and enrich the small ship features. Third, because the traditional loss function based on centroid distance and aspect ratio may result in the same loss function value for different prediction box sizes, a novel loss function based on the dual Euclidean distances of the corner point coordinates between the prediction box and the real box was proposed, which could accurately describe various overlapping box situations. Experiments using the Large-Scale SAR Ship Detection Dataset (LS-SSDD), SAR Ship Detection Dataset (SSDD), and AIR-SARShip dataset validated the efficacy and state-of-the-art performance.

Evaluating floor types during simple composting of leaf wastes

This study aimed to evaluate the type of floor in the composting process of leaf wastes. The waste consisted of cacao leaves (<em>Theobroma cacao</em> L.), rubberwood leaves (<em>Hevea brasiliensis</em>), teak leaves (<em>Tectona grandis</em> L.f), matoa or Fijian longan leaves (<em>Pometia pinnata</em>), durian leaves (<em>Durio zibethinus</em>), and grass. The research was conducted by composting the leaf mixtures in a composting box sizing 220 cm in length, 180 cm in width, and 100 cm in height. A long box was partitioned by using plastic tarpaulin into four boxes for different floors, namely cement, soil, tarpaulin, and rice husks. A mixture of leaves of 127 kg to 176 kg was introduced in each box and was composted for 145 days. Variables observed were temperature, the height of heap, the mass of material (initial and final), C/N ratio, NPK contents, compost color, odor, pH, and moisture content. The results showed that compost quality fulfilled Indonesian National Standard (SNI 19-7030-2004) regardless of the floor types. However, the type of floors affected the compost process and yield. The composting process with tarpaulin-layered or husk-layered floors produced better compost in terms of N, P, C/N ratio, and color as compared to those resulting from cemented or barely soil floors.

A Lightweight Network for Defect Detection in Nickel-Plated Punched Steel Strip Images

As a critical component of power battery, the quality of nickel-plated punched steel strip (NPSS) is closely related to the performance of battery. In practice, however, the real-time detection for defects of NPSS with limited computating resources is a challenging task. Current research on vision-based strip defect detectors emphasizes the use of large neural network models to pursue high accuracy while ignoring the training cost and hardware requirements. To accelerate the model update and deployment cycles and reduce the application cost, a lightweight detection network for accurate and fast detection of surface defects in NPSS is developed in this article. First, a new lightweight backbone network, lightweight efficient detection network (LEDNet), is designed to extract features. Depth dynamic convolution is used as the basic structure of the network to reduce the computational complexity while adaptively extracting effective features. Thereafter, a refine-residual bidirectional aggregation network is built to enhance the bidirectional propagation and reuse high- and low-level features by further optimizing and adjusting the extracted features. On this basis, feature aggregation capability is improved. Finally, the intersection-over-union k-means algorithm is used to adjust the anchor box size, which balances the significant difference in aspect ratio of surface defects of NPSS. Experimental results on the NPSS defect dataset show that our network detection accuracy can reach 87.17%, the single-sheet detection rate is 0.09 s, and the computational complexity is reduced. Compared with the state-of-the-art detectors, it achieves a strong competitive accuracy and low computational effort, which fully meets the accuracy and real-time requirements in the production environment of NPSS.

Face Recognition Model Optimization Research Based on Embedded Platform

The development of information technology has promoted the expansion of the application field of facial recognition technology. Its mainstream recognition methods rely on deep learning algorithms for calculation, but the problem of large data computation brought by its system makes it difficult to apply in embedded platform devices. As a result, this study focuses on improving recognition systems built on lightweight backend networks and builds an embedded platform system environment using multi-scale feature fusion, anchor box size optimization, the addition of channel attention mechanism weighted features, affine face alignment, and file compilation. The experimental results showed that when the number of iterations was 300, the loss value (0.46) of the improved embedded algorithm was much smaller than that of other comparison algorithms (1.42, 1.73, 2.01), and its ACC value (0.924) was significantly better than other comparison algorithms (0.915, 0.909, 0.894). The minimum system testing time consumption was 7 ms. This deep learning embedded facial recognition algorithm has high recognition accuracy and fast running speed, and is less limited by environmental conditions and data types.It is ideally suited for use in embedded hardware devices, broadening the scope of equipment matching and facial recognition algorithms’ applications. As a result, it is better suited to satisfy the demands of embedded devices and massive data processing jobs.

Bolt Positioning Detection Based on Improved YOLOv5 for Bridge Structural Health Monitoring.

To improve the stability of the bridge structure, we detect bolts in the bridge which cause the symmetry failure of the bridge center. For data acquisition, bolts are small-scale objects under complex background in images, and their feature expression ability is limited. Due to those questions, we propose a new bolt positioning detection based on improved YOLOv5 for bridge structural health monitoring. This paper makes three major contributions. Firstly, according to the calibration anchor boxes of bolts, the size and proportion parameters of the initial anchor boxes are optimized by K-means++ clustering algorithm to solve the initial clustering problem of anchor boxes in object detection. Second, the hypercolumn (HC) technique fuses the low-level global features of the trunk and the high-level local features of three different scales to solve the problem of the inefficient distribution of anchors and insufficient extraction of classification features. In this way, we improve the detection accuracy and speed of bolt detection. Finally, we establish a dataset of bridge bolts through network collection and public datasets, including 1494 images. We compare and verify the new method in the collected bolt dataset. The experimental results show that the precision (P) of the improved YOLOv5x is up to 87.3%, and the average precision (AP) is up to 86.3%, which are 6.5% and 5.9% higher than the original YOLOv5x, respectively.

ICE MAKER TOOL DESIGN

The purpose of this research is to design an ice maker using split AC machine with AutoCAD application software. This study uses a combination research design (mix methods), namely the researcher collect, analyze between quantitative and qualitative data in one study that same. Combination research design aims to complement the description of the results research on the events under study and to strengthen the research analysis. researcher choose to use explanatory which is included in the sequential model (sequence). Design This type is a mixed methods research design carried out by carrying out qualitative research first and then proceed with quantitative research. Based on the results of the design of the ice maker tool using the AutoCAD application 2016 it can be concluded that the size of the ice maker cooling box has a length of 520 mm, 460 mm wide and 520 mm high. The capacity of the number of ice cubes produced in one production process as many as 10 plastic bags of sugar one kilo. Storage box size The cooling water for the ice maker has a length of 370 mm, a width of 280 mm, and a height of 290 mm. The size of the ice maker frame or frame has a length of 1200 mm, a width of 650 mm, and has a height of 50 mm. The Split AC machine used is Split AC PK, the refrigerant that is used is refrigerant R22. Documentation of the design results of the ice maker tool using the AutoCAD application printed on A3 paper. Use accurate measuring tools so that the measurement results are more precise and precise. Research This is hoping for further research, regarding the design of the ice maker in order to obtain information that really becomes a reference for the community in good cooling. Need the existence of learning media that are validated by media experts so that when making tools they can be learning materials.

An improved faster R-CNN algorithm for assisted detection of lung nodules

The morbidity and mortality of lung cancer are increasing rapidly in every country in the world, and pulmonary nodules are the main symptoms of lung cancer in the early stage. If we can diagnose pulmonary nodules in time at the early stage and follow up and treat suspicious patients, we can effectively reduce the incidence of lung cancer. CT (Computed Tomography) has been applied to the screening of many diseases because of its high resolution. Pulmonary nodules show white round shadows in CT images. With the popularity of CT equipment, doctors need to review a large number of imaging results every day. Doctors will misjudge and miss the lesions because of reviewing CT scanning results for a long time. At this time, the method of automatic detection of pulmonary nodules by computer can relieve the pressure of doctors in reviewing CT scan results.Traditional lung nodule detection methods, such as gray threshold method and region growing method, divide the detection process into two steps: extracting candidate regions and eliminating false regions. In addition, the traditional detection method can only operate on a single image, which leads to the inability of this method to detect the batch scanning results in real time. With the continuous development of computer equipment performance and artificial intelligence, the relationship between medical image processing and deep learning is getting closer and closer. In deep learning, object detection methods such as Faster R-CNN、YOLO can complete parallel detection of batch images, and deep structure can fully extract the features of input images. Compared with traditional lung nodule detection methods, it has the characteristics of high efficiency and high precision. Faster R-CNN is a classical and high-precision two-stage object detection method. In this paper, an improved Faster R-CNN model is proposed. On the basis of Faster R-CNN, multi-scale training strategy is used to fully mine the features of different scale spaces and perform path augmentation on lower-dimensional features, which improves the small object detection ability of the model. Through Online Hard Example Mining (OHEM), the loss value is used to quantify the difficulty of candidate region detection, and the training times of the region to be detected are adaptively adjusted. Make full use of prior information to customize the size and proportion of preset boundary anchor boxes. Using deformable convolution to improve the visual field to enhance the global features and enhance the ability to extract the feature information of pulmonary nodules in the same scale space.The new model was tested on LUNA16 (Lung Nodule Analysis 2016) dataset. The detection precision of the improved Faster R-CNN model for pulmonary nodules increased from 76.4% to 90.7%, and the recall rate increased from 40.1% to 56.8% Compared with the mainstream object detection algorithms YOLOv3 and Cascade R-CNN, the improved model is superior to the above models in every index.

Multiscale anchor box and optimized classification with faster R‐CNN for object detection

AbstractFor the two‐stage object detector as a faster region‐convolutional neural network (Faster R‐CNN), upgrading the accuracy of object recognition depends on the proposal box, which is generated by the region proposal algorithms. Due to the limitations of the anchor setting of Faster RCNN, the size of the proposal box generated by the region proposal network (RPN) used is large, which would easily cause a great number of overflows in the sliding search. To improve the accuracy of object detection and remit the overflow problem of the anchor box, multi‐scale anchor box and moving overflow anchor box strategies are introduced here. Then, to increase the positive sample range of the foreground, the hierarchical weight cross entropy classification function is set for binary classification in the RPN network. These strategies could improve the accuracy of object detection. The experimental result achieves 76.2% AP on the Pascal VOC 2007(VOC 07) dataset, which is 2.7% higher than the Faster R‐CNN. The result of the Pascal VOC 2012(VOC 12) test, we achieve 75.6% AP, is improved by 2.5% compared with the Faster R‐CNN.

Helmet Wearing State Detection Based on Improved Yolov5s

At many construction sites, whether to wear a helmet is directly related to the safety of the workers. Therefore, the detection of helmet use has become a crucial monitoring tool for construction safety. However, most of the current helmet wearing detection algorithms are only dedicated to distinguishing pedestrians who wear helmets from those who do not. In order to further enrich the detection in construction scenes, this paper builds a dataset with six cases: not wearing a helmet, wearing a helmet, just wearing a hat, having a helmet, but not wearing it, wearing a helmet correctly, and wearing a helmet without wearing the chin strap. On this basis, this paper proposes a practical algorithm for detecting helmet wearing states based on the improved YOLOv5s algorithm. Firstly, according to the characteristics of the label of the dataset constructed by us, the K-means method is used to redesign the size of the prior box and match it to the corresponding feature layer to increase the accuracy of the feature extraction of the model; secondly, an additional layer is added to the algorithm to improve the ability of the model to recognize small targets; finally, the attention mechanism is introduced in the algorithm, and the CIOU_Loss function in the YOLOv5 method is replaced by the EIOU_Loss function. The experimental results indicate that the improved algorithm is more accurate than the original YOLOv5s algorithm. In addition, the finer classification also significantly enhances the detection performance of the model.

Damaged apple detection with a hybrid YOLOv3 algorithm

This paper proposes an improved You Only Look Once (YOLOv3) algorithm for automatically detecting damaged apples to promote the automation of the fruit processing industry. In the proposed method, a clustering method based on Rao-1 algorithm is introduced to optimize anchor box sizes. The clustering method uses the intersection over the union to form the objective function and the most representative anchor boxes are generated for normal and damaged apple detection. To verify the feasibility and effectiveness of the proposed method, real apple images collected from the Internet are employed. Compared with the generic YOLOv3 and Fast Region-based Convolutional Neural Network (Fast R-CNN) algorithms, the proposed method yields the highest mean average precision value for the test dataset. Therefore, it is practical to apply the proposed method for intelligent apple detection and classification tasks.

Quantitative Analysis of the Multicomponent and Spectrum-Effect Correlation of the Antispasmodic Activity of Shaoyao-Gancao Decoction.

Shaoyao-Gancao Decoction (SGD) is a well-known classic traditional Chinese medicine (TCM) with antispasmodic, anti-inflammatory, and analgesic effects. This preparation has been widely used to treat spasticity diseases in the clinic. To date, the material basis of SGD remains unclear, and the spectrum-effect correlation of its antispasmodic activity has not been reported yet. In this study, high-performance liquid chromatography (HPLC) was used to establish the fingerprint and determine the multiple components of SGD. The common peaks of fingerprints were evaluated by the similarity with the chromatographic fingerprints of the TCM. Meanwhile, the multiple components were quantified and analysed using the heatmap and box size analysis. Furthermore, data on the antispasmodic effect were extracted through in vitro smooth muscle contraction assay. Grey relational analysis combined with partial least square regression was used to study the spectrum-effect correlation of SGD. Finally, the potential antispasmolytic components were validated using an isolated tissue experiment. The HPLC fingerprint was established, and 20 common peaks were identified. The similarities of 15 batches of SGD were all above 0.965. The HPLC method for simultaneous determination of the multiple components was accurate and reliable. The contents of albiflorin, paeoniflorin, liquiritin, and glycyrrhizic acid were higher than the other components in SGD. The heatmap and box size also showed that X3 (albiflorin), X4 (paeoniflorin), X5 (liquiritin), X11 (liquirtigenin), and X16 (glycyrrhizic acid) could be used as quality indicators in the further establishment of quality standards. The spectrum-effect correlation results indicated that X4, X11, and X16 were highly correlated with antispasmolytic activity. Verification tests showed that paeoniflorin (11.7-29.25 μg/mL) and liquirtigenin (17.19-28.65 μg/mL) could significantly reduce the maximum contractile (P < 0.01). These compounds exerted concentration-dependent spasmolytic effects with the inhibitory response for acetylcholine (Ach)-evoked contraction. Thus, SGD had a significant antispasmodic effect, which resulted from the synergistic activity of its multiple components. These findings can be used for the pharmacodynamics study of SGD and are of great significance for the determination of quality markers and quality control.

Performance Evaluation of Different Object Detection Models for the Segmentation of Optical Cups and Discs.

Glaucoma is an eye disease that gradually deteriorates vision. Much research focuses on extracting information from the optic disc and optic cup, the structure used for measuring the cup-to-disc ratio. These structures are commonly segmented with deeplearning techniques, primarily using Encoder-Decoder models, which are hard to train and time-consuming. Object detection models using convolutional neural networks can extract features from fundus retinal images with good precision. However, the superiority of one model over another for a specific task is still being determined. The main goal of our approach is to compare object detection model performance to automate segment cups and discs on fundus images. This study brings the novelty of seeing the behavior of different object detection models in the detection and segmentation of the disc and the optical cup (Mask R-CNN, MS R-CNN, CARAFE, Cascade Mask R-CNN, GCNet, SOLO, Point_Rend), evaluated on Retinal Fundus Images for Glaucoma Analysis (REFUGE), and G1020 datasets. Reported metrics were Average Precision (AP), F1-score, IoU, and AUCPR. Several models achieved the highest AP with a perfect 1.000 when the threshold for IoU was set up at 0.50 on REFUGE, and the lowest was Cascade Mask R-CNN with an AP of 0.997. On the G1020 dataset, the best model was Point_Rend with an AP of 0.956, and the worst was SOLO with 0.906. It was concluded that the methods reviewed achieved excellent performance with high precision and recall values, showing efficiency and effectiveness. The problem of how many images are needed was addressed with an initial value of 100, with excellent results. Data augmentation, multi-scale handling, and anchor box size brought improvements. The capability to translate knowledge from one database to another shows promising results too.

A Multitarget Interested Region Extraction Method for Wrist X-Ray Images Based on Optimized AlexNet and Two-Class Combined Model

Bone age assessment based on X-ray Images can accurately determine the actual bone age of adolescents. Accurate extraction of the key regions of interest (ROIs) in X-ray images is required to accurately assess bone age. However, existing ROI extraction methods can only extract a few targets and have poor extraction accuracy. Thus, the strict demands for imaging in the medical field via these methods are difficult. In this article, we propose a multitarget interested region extraction method for wrist X-ray images based on optimized AlexNet and two-class combined model, named OATC, which can simultaneously extract multiple ROIs with high accuracy. Specifically, the square-wave scanning algorithm was implemented to obtain the bounding box size of each bone ROI according to the shape information of the wrist. Then, the optimized AlexNet was used to obtain the key point coordinates of each bone ROI. Bone ROIs could be extracted by combining key point coordinates with the bone bounding box size. Finally, the two-classification model was combined to improve the accuracy of ROI extraction. Experiments conducted on our wrist X-ray image dataset showed that OTAC has a fast convergence speed and small deviation. The average accuracy of extracting 14 ROIs reached 95.57%, which are 7.76% and 4.68% higher than that of VGG16 and AlexNet, respectively.