The massive influx of traffic on the Internet has made the composition of web traffic increasingly complex.Traditional port-based or protocol-based network traffic identification methods are no longer suitable for to...The massive influx of traffic on the Internet has made the composition of web traffic increasingly complex.Traditional port-based or protocol-based network traffic identification methods are no longer suitable for today’s complex and changing networks.Recently,machine learning has beenwidely applied to network traffic recognition.Still,high-dimensional features and redundant data in network traffic can lead to slow convergence problems and low identification accuracy of network traffic recognition algorithms.Taking advantage of the faster optimizationseeking capability of the jumping spider optimization algorithm(JSOA),this paper proposes a jumping spider optimization algorithmthat incorporates the harris hawk optimization(HHO)and small hole imaging(HHJSOA).We use it in network traffic identification feature selection.First,the method incorporates the HHO escape energy factor and the hard siege strategy to forma newsearch strategy for HHJSOA.This location update strategy enhances the search range of the optimal solution of HHJSOA.We use small hole imaging to update the inferior individual.Next,the feature selection problem is coded to propose a jumping spiders individual coding scheme.Multiple iterations of the HHJSOA algorithmfind the optimal individual used as the selected feature for KNN classification.Finally,we validate the classification accuracy and performance of the HHJSOA algorithm using the UNSW-NB15 dataset and KDD99 dataset.Experimental results show that compared with other algorithms for the UNSW-NB15 dataset,the improvement is at least 0.0705,0.00147,and 1 on the accuracy,fitness value,and the number of features.In addition,compared with other feature selectionmethods for the same datasets,the proposed algorithmhas faster convergence,better merit-seeking,and robustness.Therefore,HHJSOAcan improve the classification accuracy and solve the problem that the network traffic recognition algorithm needs to be faster to converge and easily fall into local optimum due to high-dimensional features.展开更多
Small cell carcinoma of cervix (SCCC) is a rare disease with highly aggressive behaviour and is pathologically hard to diagnose.In this study, the clinicopathological features, diagnosis, treatment and prognosis of th...Small cell carcinoma of cervix (SCCC) is a rare disease with highly aggressive behaviour and is pathologically hard to diagnose.In this study, the clinicopathological features, diagnosis, treatment and prognosis of the condition were examined.Clinical records and follow-up data of 7 cases of SCCC were retrospectively studied.Our results showed that five non-recurrent cases initially presented irregular vaginal bleeding or increased apocenosis of varying degrees.Pathological examination revealed that the stroma was diffusely infiltrated with small monomorphous cells ranging from round to oval shape.Three cases were immunohistochemically confirmed.One case was accompanied with squamous cell cancer.Of the 7 cases, one case was classified as stage Ⅰb 1, two stageⅠ b2, one stage Ⅱ a, one stage Ⅱb , and one stage Ⅲ b.On the basis of their stages of condition, one subject with stage III b underwent chemotherapy, and one with stage Ib2 received extensive hysterectomy plus pelvic lymphadenectomy, while the other 5 cases were treated by extensive hysterectomy and pelvic lymphadenectomy in combination with pre-and/or post-operative adjuvant chemotherapy and radiotherapy.Of the 7 patients, 4 had relapse-free survival of 14, 14, 16 and 28 months respectively.It is concluded that SCCC is an aggressive tumor with propensity for early pelvis lymph node metastases.Early-stage patients should be treated by extensive hysterectomy and pelvic lymphadenectomy in combination with pre-and/or post-operative adjuvant chemotherapy and radiotherapy.展开更多
Accessibility and flexibility of small roadside parks make them significant transitional spaces in urban landscape environment. Three representative small parks in Zhanjiang City, a typical tropical city in south Chin...Accessibility and flexibility of small roadside parks make them significant transitional spaces in urban landscape environment. Three representative small parks in Zhanjiang City, a typical tropical city in south China, were selected to analyze their location features, spatial processing, demonstration of regional landscapes and recreational characteristics. It was proposed that construction of small roadside parks in tropical area should put human needs on the priority, present regional features of tropical garden landscapes, and focus on inheritance and innovation of regional cultures.展开更多
To investigate the clinical and computed tomography(CT)features of desmoplastic small round cell tumor(DSRCT),we retrospectively analyzed the clinical presentations,treatment and outcome,as well as CT manifestations o...To investigate the clinical and computed tomography(CT)features of desmoplastic small round cell tumor(DSRCT),we retrospectively analyzed the clinical presentations,treatment and outcome,as well as CT manifestations of four cases of DSRCT confirmed by surgery and pathology.The CT manifestations of DSRCT were as follows:(1)multiple soft-tissue masses or diffuse peritoneal thickening in the abdomen and pelvis,with the dominant mass usually located in the pelvic cavity;(2)masses without an apparent organbased primary site;(3)mild to moderate homogeneous or heterogeneous enhancement in solid area on enhanced CT;and(4)secondary manifestations,such as ascites,hepatic metastases,lymphadenopathy,hydronephrosis and hydroureter.The prognosis and overall survival rates were generally poor.Commonly used treatment strategies including aggressive tumor resection,polychemotherapy,and radiotherapy,showed various therapeutic effects.CT of DSRCT shows characteristic features that are helpful in diagnosis.Early discovery and complete resection,coupled with postoperative adjuvant chemotherapy,are important for prognosis of DSRCT.Whole abdominopelvic rather than locoregional radiotherapy is more effective for unresectable DSRCT.展开更多
Object detection has been studied for many years.The convolutional neural network has made great progress in the accuracy and speed of object detection.However,due to the low resolution of small objects and the repres...Object detection has been studied for many years.The convolutional neural network has made great progress in the accuracy and speed of object detection.However,due to the low resolution of small objects and the representation of fuzzy features,one of the challenges now is how to effectively detect small objects in images.Existing target detectors for small objects:one is to use high-resolution images as input,the other is to increase the depth of the CNN network,but these two methods will undoubtedly increase the cost of calculation and time-consuming.In this paper,based on the RefineDet network framework,we propose our network structure RF2Det by introducing Receptive Field Block to solve the problem of small object detection,so as to achieve the balance of speed and accuracy.At the same time,we propose a Medium-level Feature Pyramid Networks,which combines appropriate high-level context features with low-level features,so that the network can use the features of both the low-level and the high-level for multi-scale target detection,and the accuracy of the small target detection task based on the low-level features is improved.Extensive experiments on the MS COCO dataset demonstrate that compared to other most advanced methods,our proposed method shows significant performance improvement in the detection of small objects.展开更多
Objective:To study the relevance of EGFR gene mutation with pathological features and prognosis in patients with non-small-cell lung carcinoma.Methods:A total of 297 patients from July 2009 to May 2013 were chosen as ...Objective:To study the relevance of EGFR gene mutation with pathological features and prognosis in patients with non-small-cell lung carcinoma.Methods:A total of 297 patients from July 2009 to May 2013 were chosen as objects.EGFR gene mutation were detected with fluorescence quantitative PCR.Relevance of EGFR gene mutation with clinical and pathological features was analyzed,and the prognosis of EGFR- mutant-patients and that of EGFR- wide type-patients was compared.Results:In 297 patients.136(45.79%) showed EGFR gene mutation.EGFR gene mutation had no significant relevance with age.gender,smoking history,family history of cancer and clinical stage(P>0.05);there was significant relevance between EGFR gene mutation and blood type,pathologic types,differentiation and diameter of cancer(P<0.05).The difference between prognosis of EGFR- mutant-patients and that of EGFR- wide type-patients was statistical significance(P<0.05).Conclusions:EGFR gene mutation has significant relevance with pathological features,the prognosis of EGFRmutant-paticnts is better than that of EGFR- wide type-patients.展开更多
A large database is desired for machine learning(ML) technology to make accurate predictions of materials physicochemical properties based on their molecular structure.When a large database is not available,the develo...A large database is desired for machine learning(ML) technology to make accurate predictions of materials physicochemical properties based on their molecular structure.When a large database is not available,the development of proper featurization method based on physicochemical nature of target proprieties can improve the predictive power of ML models with a smaller database.In this work,we show that two new featurization methods,volume occupation spatial matrix and heat contribution spatial matrix,can improve the accuracy in predicting energetic materials' crystal density(ρ_(crystal)) and solid phase enthalpy of formation(H_(f,solid)) using a database containing 451 energetic molecules.Their mean absolute errors are reduced from 0.048 g/cm~3 and 24.67 kcal/mol to 0.035 g/cm~3 and 9.66 kcal/mol,respectively.By leave-one-out-cross-validation,the newly developed ML models can be used to determine the performance of most kinds of energetic materials except cubanes.Our ML models are applied to predict ρ_(crystal) and H_(f,solid) of CHON-based molecules of the 150 million sized PubChem database,and screened out 56 candidates with competitive detonation performance and reasonable chemical structures.With further improvement in future,spatial matrices have the potential of becoming multifunctional ML simulation tools that could provide even better predictions in wider fields of materials science.展开更多
The four nucleotides (bases), A, T (U), G and C in small genomes, virus DNA/RNA, organelle and plastid genomes were also arranged sophisticatedly in the structural features in a single-strand with 1) reverse-complemen...The four nucleotides (bases), A, T (U), G and C in small genomes, virus DNA/RNA, organelle and plastid genomes were also arranged sophisticatedly in the structural features in a single-strand with 1) reverse-complement symmetry of base or base sequences, 2) bias of four bases, 3) multiple fractality of the distribution of each four bases depending on the distance in double logarithmic plot (power spectrum) of L (the distance of a base to the next base) vs. P (L) (the probability of the base-distribution at L), although their genomes were composed of low numbers of the four bases, and the base-symmetry was rather lower than the prokaryotic-and the eukaryotic cells. In the case of the genomic DNA composed of less than 10,000 nt, it was better than to be partitioned at 10 of the L-value, and the structural features for the biologically active genomic DNA were observed as the large genomes. As the results, the base sequences of the genomic DNA including the genomic-RNA might be universal in all genomes. In addition, the relationship between the structural features of the genome and the biological complexity was discussed.展开更多
BACKGROUND Angiosarcoma is a highly malignant soft-tissue sarcoma derived from vascular endothelial cells that mainly occurs in the skin and subcutaneous tissues.Smallintestinal angiosarcomas are rare,and the prognosi...BACKGROUND Angiosarcoma is a highly malignant soft-tissue sarcoma derived from vascular endothelial cells that mainly occurs in the skin and subcutaneous tissues.Smallintestinal angiosarcomas are rare,and the prognosis is poor.CASE SUMMARY We reported a case of primary multifocal ileal angiosarcoma and analyze previously reported cases to improve our understanding of small intestinal angiosarcoma.Small intestinal angiosarcoma is more common in elderly and male patients.Gastrointestinal bleeding,anemia,abdominal pain,weakness,and weight loss were the common symptoms.CD31,CD34,factor VIII-related antigen,ETS-related gene,friend leukemia integration 1,and von Willebrand factor are valuable immunohistochemical markers for the diagnosis of small-intestinal angiosarcoma.Small-intestinal angiosarcoma most commonly occurs in the jejunum,followed by the ileum and duodenum.Radiation and toxicant exposure are risk factors for angiosarcoma.After a definite diagnosis,the mean and median survival time was 8 mo and 3 mo,respectively.Kaplan-Meier survival analysis showed that age,infiltration depth,chemotherapy,and the number of small intestinal segments invaded by tumor lesions were prognostic factors for small intestinal angiosarcoma.Multivariate Cox regression analysis showed that chemotherapy and surgery significantly improved patient prognosis.CONCLUSION Angiosarcoma should be considered for unexplained melena and abdominal pain,especially in older men and patients with a history of radiation exposure.Prompt treatment,including surgery and adjuvant chemotherapy,is essential to prolonging patient survival.展开更多
This paper proposes a real-time detection method to improve the Infrared small target detection CenterNet(ISTD-CenterNet)network for detecting small infrared targets in complex environments.The method eliminates the n...This paper proposes a real-time detection method to improve the Infrared small target detection CenterNet(ISTD-CenterNet)network for detecting small infrared targets in complex environments.The method eliminates the need for an anchor frame,addressing the issues of low accuracy and slow speed.HRNet is used as the framework for feature extraction,and an ECBAM attention module is added to each stage branch for intelligent identification of the positions of small targets and significant objects.A scale enhancement module is also added to obtain a high-level semantic representation and fine-resolution prediction map for the entire infrared image.Besides,an improved sensory field enhancement module is designed to leverage semantic information in low-resolution feature maps,and a convolutional attention mechanism module is used to increase network stability and convergence speed.Comparison experiments conducted on the infrared small target data set ESIRST.The experiments show that compared to the benchmark network CenterNet-HRNet,the proposed ISTD-CenterNet improves the recall by 22.85%and the detection accuracy by 13.36%.Compared to the state-of-the-art YOLOv5small,the ISTD-CenterNet recall is improved by 5.88%,the detection precision is improved by 2.33%,and the detection frame rate is 48.94 frames/sec,which realizes the accurate real-time detection of small infrared targets.展开更多
无人机高空航拍图像中车辆像素占比极低,目标可视化信息较少,在目标检测任务中容易漏检和误检。因此,本文提出一种基于改进YOLOX(You Only Look Once X)的无人机高空航拍视角下小尺度车辆精确检测方法。首先,为增强网络对低级特征的提...无人机高空航拍图像中车辆像素占比极低,目标可视化信息较少,在目标检测任务中容易漏检和误检。因此,本文提出一种基于改进YOLOX(You Only Look Once X)的无人机高空航拍视角下小尺度车辆精确检测方法。首先,为增强网络对低级特征的提取能力,在原始YOLOX预测头部增加一个160 pixel×160 pixel的浅层特征提取网络;其次,在骨干网络后端嵌入基于归一化的注意力机制模块(Normalization-based Attention Module,NAM),以抑制冗余的非显著特征表达;最后,为了增大小尺度车辆的相对像素比,提升网络捕捉有效特征信息的能力,提出一种基于滑动窗口的图像切分检测方法。试验结果表明,改进YOLOX网络表现出良好的检测效能,检测精度达到了84.58%,优于典型的目标检测网络Faster R-CNN(79.95%)、YOLOv3(83.69%)、YOLOv5(84.31%)及YOLOX(83.10%)。此外,改进YOLOX能够有效解决无人机高空航拍图像中小尺度车辆的漏检和误检问题,且预测框更贴合车辆的实际轮廓;同时,在不同航拍高度的目标检测任务中具有较高的鲁棒性。展开更多
基金funded by the National Natural Science Foundation of China under Grant No.61602162.
文摘The massive influx of traffic on the Internet has made the composition of web traffic increasingly complex.Traditional port-based or protocol-based network traffic identification methods are no longer suitable for today’s complex and changing networks.Recently,machine learning has beenwidely applied to network traffic recognition.Still,high-dimensional features and redundant data in network traffic can lead to slow convergence problems and low identification accuracy of network traffic recognition algorithms.Taking advantage of the faster optimizationseeking capability of the jumping spider optimization algorithm(JSOA),this paper proposes a jumping spider optimization algorithmthat incorporates the harris hawk optimization(HHO)and small hole imaging(HHJSOA).We use it in network traffic identification feature selection.First,the method incorporates the HHO escape energy factor and the hard siege strategy to forma newsearch strategy for HHJSOA.This location update strategy enhances the search range of the optimal solution of HHJSOA.We use small hole imaging to update the inferior individual.Next,the feature selection problem is coded to propose a jumping spiders individual coding scheme.Multiple iterations of the HHJSOA algorithmfind the optimal individual used as the selected feature for KNN classification.Finally,we validate the classification accuracy and performance of the HHJSOA algorithm using the UNSW-NB15 dataset and KDD99 dataset.Experimental results show that compared with other algorithms for the UNSW-NB15 dataset,the improvement is at least 0.0705,0.00147,and 1 on the accuracy,fitness value,and the number of features.In addition,compared with other feature selectionmethods for the same datasets,the proposed algorithmhas faster convergence,better merit-seeking,and robustness.Therefore,HHJSOAcan improve the classification accuracy and solve the problem that the network traffic recognition algorithm needs to be faster to converge and easily fall into local optimum due to high-dimensional features.
基金supported by a grant from the Program of Scientific Innovation of Huazhong University of Science and Technology (No.HF-05-035-07-540)
文摘Small cell carcinoma of cervix (SCCC) is a rare disease with highly aggressive behaviour and is pathologically hard to diagnose.In this study, the clinicopathological features, diagnosis, treatment and prognosis of the condition were examined.Clinical records and follow-up data of 7 cases of SCCC were retrospectively studied.Our results showed that five non-recurrent cases initially presented irregular vaginal bleeding or increased apocenosis of varying degrees.Pathological examination revealed that the stroma was diffusely infiltrated with small monomorphous cells ranging from round to oval shape.Three cases were immunohistochemically confirmed.One case was accompanied with squamous cell cancer.Of the 7 cases, one case was classified as stage Ⅰb 1, two stageⅠ b2, one stage Ⅱ a, one stage Ⅱb , and one stage Ⅲ b.On the basis of their stages of condition, one subject with stage III b underwent chemotherapy, and one with stage Ib2 received extensive hysterectomy plus pelvic lymphadenectomy, while the other 5 cases were treated by extensive hysterectomy and pelvic lymphadenectomy in combination with pre-and/or post-operative adjuvant chemotherapy and radiotherapy.Of the 7 patients, 4 had relapse-free survival of 14, 14, 16 and 28 months respectively.It is concluded that SCCC is an aggressive tumor with propensity for early pelvis lymph node metastases.Early-stage patients should be treated by extensive hysterectomy and pelvic lymphadenectomy in combination with pre-and/or post-operative adjuvant chemotherapy and radiotherapy.
基金Supported by China National Natural Science Fund(51208118)
文摘Accessibility and flexibility of small roadside parks make them significant transitional spaces in urban landscape environment. Three representative small parks in Zhanjiang City, a typical tropical city in south China, were selected to analyze their location features, spatial processing, demonstration of regional landscapes and recreational characteristics. It was proposed that construction of small roadside parks in tropical area should put human needs on the priority, present regional features of tropical garden landscapes, and focus on inheritance and innovation of regional cultures.
文摘To investigate the clinical and computed tomography(CT)features of desmoplastic small round cell tumor(DSRCT),we retrospectively analyzed the clinical presentations,treatment and outcome,as well as CT manifestations of four cases of DSRCT confirmed by surgery and pathology.The CT manifestations of DSRCT were as follows:(1)multiple soft-tissue masses or diffuse peritoneal thickening in the abdomen and pelvis,with the dominant mass usually located in the pelvic cavity;(2)masses without an apparent organbased primary site;(3)mild to moderate homogeneous or heterogeneous enhancement in solid area on enhanced CT;and(4)secondary manifestations,such as ascites,hepatic metastases,lymphadenopathy,hydronephrosis and hydroureter.The prognosis and overall survival rates were generally poor.Commonly used treatment strategies including aggressive tumor resection,polychemotherapy,and radiotherapy,showed various therapeutic effects.CT of DSRCT shows characteristic features that are helpful in diagnosis.Early discovery and complete resection,coupled with postoperative adjuvant chemotherapy,are important for prognosis of DSRCT.Whole abdominopelvic rather than locoregional radiotherapy is more effective for unresectable DSRCT.
文摘Object detection has been studied for many years.The convolutional neural network has made great progress in the accuracy and speed of object detection.However,due to the low resolution of small objects and the representation of fuzzy features,one of the challenges now is how to effectively detect small objects in images.Existing target detectors for small objects:one is to use high-resolution images as input,the other is to increase the depth of the CNN network,but these two methods will undoubtedly increase the cost of calculation and time-consuming.In this paper,based on the RefineDet network framework,we propose our network structure RF2Det by introducing Receptive Field Block to solve the problem of small object detection,so as to achieve the balance of speed and accuracy.At the same time,we propose a Medium-level Feature Pyramid Networks,which combines appropriate high-level context features with low-level features,so that the network can use the features of both the low-level and the high-level for multi-scale target detection,and the accuracy of the small target detection task based on the low-level features is improved.Extensive experiments on the MS COCO dataset demonstrate that compared to other most advanced methods,our proposed method shows significant performance improvement in the detection of small objects.
基金supported by Project Development Plan of Yantai city Science and Technology(No.2013WS229)
文摘Objective:To study the relevance of EGFR gene mutation with pathological features and prognosis in patients with non-small-cell lung carcinoma.Methods:A total of 297 patients from July 2009 to May 2013 were chosen as objects.EGFR gene mutation were detected with fluorescence quantitative PCR.Relevance of EGFR gene mutation with clinical and pathological features was analyzed,and the prognosis of EGFR- mutant-patients and that of EGFR- wide type-patients was compared.Results:In 297 patients.136(45.79%) showed EGFR gene mutation.EGFR gene mutation had no significant relevance with age.gender,smoking history,family history of cancer and clinical stage(P>0.05);there was significant relevance between EGFR gene mutation and blood type,pathologic types,differentiation and diameter of cancer(P<0.05).The difference between prognosis of EGFR- mutant-patients and that of EGFR- wide type-patients was statistical significance(P<0.05).Conclusions:EGFR gene mutation has significant relevance with pathological features,the prognosis of EGFRmutant-paticnts is better than that of EGFR- wide type-patients.
基金support from the Ministry of Education(MOE) Singapore Tier 1 (RG8/20)。
文摘A large database is desired for machine learning(ML) technology to make accurate predictions of materials physicochemical properties based on their molecular structure.When a large database is not available,the development of proper featurization method based on physicochemical nature of target proprieties can improve the predictive power of ML models with a smaller database.In this work,we show that two new featurization methods,volume occupation spatial matrix and heat contribution spatial matrix,can improve the accuracy in predicting energetic materials' crystal density(ρ_(crystal)) and solid phase enthalpy of formation(H_(f,solid)) using a database containing 451 energetic molecules.Their mean absolute errors are reduced from 0.048 g/cm~3 and 24.67 kcal/mol to 0.035 g/cm~3 and 9.66 kcal/mol,respectively.By leave-one-out-cross-validation,the newly developed ML models can be used to determine the performance of most kinds of energetic materials except cubanes.Our ML models are applied to predict ρ_(crystal) and H_(f,solid) of CHON-based molecules of the 150 million sized PubChem database,and screened out 56 candidates with competitive detonation performance and reasonable chemical structures.With further improvement in future,spatial matrices have the potential of becoming multifunctional ML simulation tools that could provide even better predictions in wider fields of materials science.
文摘The four nucleotides (bases), A, T (U), G and C in small genomes, virus DNA/RNA, organelle and plastid genomes were also arranged sophisticatedly in the structural features in a single-strand with 1) reverse-complement symmetry of base or base sequences, 2) bias of four bases, 3) multiple fractality of the distribution of each four bases depending on the distance in double logarithmic plot (power spectrum) of L (the distance of a base to the next base) vs. P (L) (the probability of the base-distribution at L), although their genomes were composed of low numbers of the four bases, and the base-symmetry was rather lower than the prokaryotic-and the eukaryotic cells. In the case of the genomic DNA composed of less than 10,000 nt, it was better than to be partitioned at 10 of the L-value, and the structural features for the biologically active genomic DNA were observed as the large genomes. As the results, the base sequences of the genomic DNA including the genomic-RNA might be universal in all genomes. In addition, the relationship between the structural features of the genome and the biological complexity was discussed.
文摘BACKGROUND Angiosarcoma is a highly malignant soft-tissue sarcoma derived from vascular endothelial cells that mainly occurs in the skin and subcutaneous tissues.Smallintestinal angiosarcomas are rare,and the prognosis is poor.CASE SUMMARY We reported a case of primary multifocal ileal angiosarcoma and analyze previously reported cases to improve our understanding of small intestinal angiosarcoma.Small intestinal angiosarcoma is more common in elderly and male patients.Gastrointestinal bleeding,anemia,abdominal pain,weakness,and weight loss were the common symptoms.CD31,CD34,factor VIII-related antigen,ETS-related gene,friend leukemia integration 1,and von Willebrand factor are valuable immunohistochemical markers for the diagnosis of small-intestinal angiosarcoma.Small-intestinal angiosarcoma most commonly occurs in the jejunum,followed by the ileum and duodenum.Radiation and toxicant exposure are risk factors for angiosarcoma.After a definite diagnosis,the mean and median survival time was 8 mo and 3 mo,respectively.Kaplan-Meier survival analysis showed that age,infiltration depth,chemotherapy,and the number of small intestinal segments invaded by tumor lesions were prognostic factors for small intestinal angiosarcoma.Multivariate Cox regression analysis showed that chemotherapy and surgery significantly improved patient prognosis.CONCLUSION Angiosarcoma should be considered for unexplained melena and abdominal pain,especially in older men and patients with a history of radiation exposure.Prompt treatment,including surgery and adjuvant chemotherapy,is essential to prolonging patient survival.
基金funded by National Natural Science Foundation of China,Fund Number 61703424.
文摘This paper proposes a real-time detection method to improve the Infrared small target detection CenterNet(ISTD-CenterNet)network for detecting small infrared targets in complex environments.The method eliminates the need for an anchor frame,addressing the issues of low accuracy and slow speed.HRNet is used as the framework for feature extraction,and an ECBAM attention module is added to each stage branch for intelligent identification of the positions of small targets and significant objects.A scale enhancement module is also added to obtain a high-level semantic representation and fine-resolution prediction map for the entire infrared image.Besides,an improved sensory field enhancement module is designed to leverage semantic information in low-resolution feature maps,and a convolutional attention mechanism module is used to increase network stability and convergence speed.Comparison experiments conducted on the infrared small target data set ESIRST.The experiments show that compared to the benchmark network CenterNet-HRNet,the proposed ISTD-CenterNet improves the recall by 22.85%and the detection accuracy by 13.36%.Compared to the state-of-the-art YOLOv5small,the ISTD-CenterNet recall is improved by 5.88%,the detection precision is improved by 2.33%,and the detection frame rate is 48.94 frames/sec,which realizes the accurate real-time detection of small infrared targets.
文摘无人机高空航拍图像中车辆像素占比极低,目标可视化信息较少,在目标检测任务中容易漏检和误检。因此,本文提出一种基于改进YOLOX(You Only Look Once X)的无人机高空航拍视角下小尺度车辆精确检测方法。首先,为增强网络对低级特征的提取能力,在原始YOLOX预测头部增加一个160 pixel×160 pixel的浅层特征提取网络;其次,在骨干网络后端嵌入基于归一化的注意力机制模块(Normalization-based Attention Module,NAM),以抑制冗余的非显著特征表达;最后,为了增大小尺度车辆的相对像素比,提升网络捕捉有效特征信息的能力,提出一种基于滑动窗口的图像切分检测方法。试验结果表明,改进YOLOX网络表现出良好的检测效能,检测精度达到了84.58%,优于典型的目标检测网络Faster R-CNN(79.95%)、YOLOv3(83.69%)、YOLOv5(84.31%)及YOLOX(83.10%)。此外,改进YOLOX能够有效解决无人机高空航拍图像中小尺度车辆的漏检和误检问题,且预测框更贴合车辆的实际轮廓;同时,在不同航拍高度的目标检测任务中具有较高的鲁棒性。