The objective of style transfer is to maintain the content of an image while transferring the style of another image.However,conventional methods face challenges in preserving facial features,especially in Korean port...The objective of style transfer is to maintain the content of an image while transferring the style of another image.However,conventional methods face challenges in preserving facial features,especially in Korean portraits where elements like the“Gat”(a traditional Korean hat)are prevalent.This paper proposes a deep learning network designed to perform style transfer that includes the“Gat”while preserving the identity of the face.Unlike traditional style transfer techniques,the proposed method aims to preserve the texture,attire,and the“Gat”in the style image by employing image sharpening and face landmark,with the GAN.The color,texture,and intensity were extracted differently based on the characteristics of each block and layer of the pre-trained VGG-16,and only the necessary elements during training were preserved using a facial landmark mask.The head area was presented using the eyebrow area to transfer the“Gat”.Furthermore,the identity of the face was retained,and style correlation was considered based on the Gram matrix.To evaluate performance,we introduced a metric using PSNR and SSIM,with an emphasis on median values through new weightings for style transfer in Korean portraits.Additionally,we have conducted a survey that evaluated the content,style,and naturalness of the transferred results,and based on the assessment,we can confidently conclude that our method to maintain the integrity of content surpasses the previous research.Our approach,enriched by landmarks preservation and diverse loss functions,including those related to“Gat”,outperformed previous researches in facial identity preservation.展开更多
Imagery assessment is an efficient method for detecting craniofacial anomalies.A cephalometric landmark matching approach may help in orthodontic diagnosis,craniofacial growth assessment and treatment planning.Automati...Imagery assessment is an efficient method for detecting craniofacial anomalies.A cephalometric landmark matching approach may help in orthodontic diagnosis,craniofacial growth assessment and treatment planning.Automatic landmark matching and anomalies detection helps face the manual labelling lim-itations and optimize preoperative planning of maxillofacial surgery.The aim of this study was to develop an accurate Cephalometric Landmark Matching method as well as an automatic system for anatomical anomalies classification.First,the Active Appearance Model(AAM)was used for the matching process.This pro-cess was achieved by the Ant Colony Optimization(ACO)algorithm enriched with proximity information.Then,the maxillofacial anomalies were classified using the Support Vector Machine(SVM).The experiments were conducted on X-ray cephalograms of 400 patients where the ground truth was produced by two experts.The frameworks achieved a landmark matching error(LE)of 0.50±1.04 and a successful landmark matching of 89.47%in the 2 mm and 3 mm range and of 100%in the 4 mm range.The classification of anomalies achieved an accuracy of 98.75%.Compared to previous work,the proposed approach is simpler and has a comparable range of acceptable matching cost and anomaly classification.Results have also shown that it outperformed the K-nearest neigh-bors(KNN)classifier.展开更多
The facial landmarks can provide valuable information for expression-related tasks.However,most approaches only use landmarks for segmentation preprocessing or directly input them into the neural network for fully con...The facial landmarks can provide valuable information for expression-related tasks.However,most approaches only use landmarks for segmentation preprocessing or directly input them into the neural network for fully connection.Such simple combination not only fails to pass the spatial information to network,but also increases calculation amounts.The method proposed in this paper aims to integrate facial landmarks-driven representation into the triplet network.The spatial information provided by landmarks is introduced into the feature extraction process,so that the model can better capture the location relationship.In addition,coordinate information is also integrated into the triple loss calculation to further enhance similarity prediction.Specifically,for each image,the coordinates of 68 landmarks are detected,and then a region attention map based on these landmarks is generated.For the feature map output by the shallow convolutional layer,it will be multiplied with the attention map to correct the feature activation,so as to strengthen the key region and weaken the unimportant region.Finally,the optimized embedding output can be further used for downstream tasks.Three embeddings of three images output by the network can be regarded as a triplet representation for similarity computation.Through the CK+dataset,the effectiveness of such an optimized feature extraction is verified.After that,it is applied to facial expression similarity tasks.The results on the facial expression comparison(FEC)dataset show that the accuracy rate will be significantly improved after the landmark information is introduced.展开更多
Australian Wool Innovation (AWI) today announced it had engaged world renowned style and innovation agency PeclersParis to provide colour and design trend direction for the Woolmark.Peclers’ team of fashion experts...Australian Wool Innovation (AWI) today announced it had engaged world renowned style and innovation agency PeclersParis to provide colour and design trend direction for the Woolmark.Peclers’ team of fashion experts and trend forecasters this week展开更多
New studies suggest that the Kaili Biota in southwestern China's Guizhou Province could be the third revealing evidence for the spectacular Cambrian Explosion of life about 530 million years ago, next to the Burge...New studies suggest that the Kaili Biota in southwestern China's Guizhou Province could be the third revealing evidence for the spectacular Cambrian Explosion of life about 530 million years ago, next to the Burgess Shale Biota of Canada, found in 1909,and the Chengjiang Biota, brought to light in Yunnan Province, China,in 1984.展开更多
The accuracy and repeatability of computer aided cervical vertebra landmarking (CACVL) were investigated in cephalogram.120 adolescents (60 boys,60 girls) aged from 9.1 to 17.2 years old were randomly selected.Twenty-...The accuracy and repeatability of computer aided cervical vertebra landmarking (CACVL) were investigated in cephalogram.120 adolescents (60 boys,60 girls) aged from 9.1 to 17.2 years old were randomly selected.Twenty-seven landmarks from the second to fifth cervical vertebrae on the lat-eral cephalogram were identified.In this study,the system of CACVL was developed and used to iden-tify and calculate the landmarks by fast marching method and parabolic curve fitting.The accuracy and repeatability in CACVL group were compared with those in two manual landmarking groups [orthodon-tic experts (OE) group and orthodontic novices (ON) group].The results showed that,as for the accu-racy,there was no significant difference between CACVL group and OE group no matter in x-axis or y-axis (P>0.05),but there was significant difference between CACVL group and ON group,as well as OE group and ON group in both axes (P<0.05).As for the repeatability,CACVL group was more reli-able than OE group and ON group in both axes.It is concluded that CACVL has the same or higher ac-curacy,better repeatability and less workload than manual landmarking methods.It’s reliable for cervi-cal parameters identification on the lateral cephalogram and cervical vertebral maturation prediction in orthodontic practice and research.展开更多
Despite extensive investigations,no precursor patterns for reliably predicting major earthquakes have thus far been identified.Seismogenic locked segments that can accumulate adequate strain energy to cause major eart...Despite extensive investigations,no precursor patterns for reliably predicting major earthquakes have thus far been identified.Seismogenic locked segments that can accumulate adequate strain energy to cause major earthquakes are highly heterogeneous and low brittle.The progressive cracking of the locked segments with these properties can produce an interesting seismic phenomenon:a landmark earthquake and a sequence of smaller subsequent earthquakes(pre-shocks)always arise prior to another landmark earthquake within a well-defined seismic zone and its current seismic period.Applying a mechanical model,magnitude constraint conditions,and case study data of 62 worldwide seismic zones,we show that two adjacent landmark earthquakes reliably occur at the volume-expansion point and peak-stress point(rupture)of a locked segment;thus,the former is an identified precursor for the latter.Such a precursor seismicity pattern before the locked-segment rupture has definite physical meanings,and it is universal regardless of the focal depth.Because the evolution of landmark earthquakes follows a deterministic rule described by the model,they are predictable.The results of this study lay a firm physical foundation for reliably predicting the occurrence of future landmark earthquakes in a seismic zone and can greatly improve our understanding of earthquake generation mechanism.展开更多
High-density street-level reliable landmarks are one of the important foundations for street-level geolocation.However,the existing methods cannot obtain enough street-level landmarks in a short period of time.In this...High-density street-level reliable landmarks are one of the important foundations for street-level geolocation.However,the existing methods cannot obtain enough street-level landmarks in a short period of time.In this paper,a street-level landmarks acquisition method based on SVM(Support Vector Machine)classifiers is proposed.Firstly,the port detection results of IPs with known services are vectorized,and the vectorization results are used as an input of the SVM training.Then,the kernel function and penalty factor are adjusted for SVM classifiers training,and the optimal SVM classifiers are obtained.After that,the classifier sequence is constructed,and the IPs with unknown service are classified using the sequence.Finally,according to the domain name corresponding to the IP,the relationship between the classified server IP and organization name is established.The experimental results in Guangzhou and Wuhan city in China show that the proposed method can be as a supplement to existing typical methods since the number of obtained street-level landmarks is increased substantially,and the median geolocation error using evaluated landmarks is reduced by about 2 km.展开更多
AIM To characterize esophageal endoluminal landmarks to permit radial and longitudinal esophageal orientation and accurate lesion location.METHODS Distance from the incisors and radial orientation were estimated for t...AIM To characterize esophageal endoluminal landmarks to permit radial and longitudinal esophageal orientation and accurate lesion location.METHODS Distance from the incisors and radial orientation were estimated for the main left bronchus and the left atrium landmarks in 207 consecutive patients using white light examination. A sub-study was also performed using white light followed by endoscopic ultrasound(EUS) in 25 consecutive patients to confirm the findings.The scope orientation throughout the exam was maintained at the natural axis,where the left esophageal quadrant corresponds to the area between 6 and 9 o'clock. When an anatomical landmark was identified, it was recorded with a photograph and its quadrant orientation and distance from the incisors were determined. The reference points to obtain the distances and radial orientation were as follows: the midpoint of the left main bronchus and the most intense pulsatile zone of the left atrium. With the video processor system set to moderate insufflation, measurements were obtained at the end of the patients' air expiration.RESULTS The left main bronchus and left atrium esophageal landmarks were identified using white light in 99% and 100% of subjects at a mean distance of 25.8 cm(SD2.3), and 31.4 cm(SD 2.4) from the incisors, respectively. The left main bronchus landmark was found to be a tubular, concave, non-pulsatile, esophageal external compression, occupying approximately 1/4 of the circumference. The left atrium landmark was identified as a round, convex, pulsatile, esophageal external compression, occupying approximately 1/4 of the circumference. Both landmarks were identified using white light on the anterior esophageal quadrant. In the substudy, the left main bronchus was identified in 24(92%) patients at 25.4 cm(SD2.1) and 26.7 cm(SD 1.9) from the incisors, by white light and EUS, respectively.The left atrium was recognized in all patients at 30.5 cm(SD 1.9), and 31.6 cm(SD2.3) from the incisors, by both white light and EUS, respectively. EUS confirmed that the landmarks corresponded to these two structures, respectively, and that they were located on the anterior esophageal wall. The Bland-Altman plot demonstrated high agreement between the white light and EUS measurements.CONCLUSION This study provides an endoscopic characterization of esophageal landmarks corresponding to the left main bronchus and left atrium, to permit radial and longitudinal orientation and accurate lesion location.展开更多
The effectiveness of mobile robot aided for architectural construction depends strongly on its accurate localization ability.Localization of mobile robot is increasingly important for the printing of buildings in the ...The effectiveness of mobile robot aided for architectural construction depends strongly on its accurate localization ability.Localization of mobile robot is increasingly important for the printing of buildings in the construction scene.Although many available studies on the localization have been conducted,only a few studies have addressed the more challenging problem of localization for mobile robot in large-scale ongoing and featureless scenes.To realize the accurate localization of mobile robot in designated stations,we build an artificial landmark map and propose a novel nonlinear optimization algorithm based on graphs to reduce the uncertainty of the whole map.Then,the performances of localization for mobile robot based on the original and optimized map are compared and evaluated.Finally,experimental results show that the average absolute localization errors that adopted the proposed algorithm is reduced by about 21%compared to that of the original map.展开更多
Existing IP geolocation algorithms based on delay similarity often rely on the principle that geographically adjacent IPs have similar delays.However,this principle is often invalid in real Internet environment,which ...Existing IP geolocation algorithms based on delay similarity often rely on the principle that geographically adjacent IPs have similar delays.However,this principle is often invalid in real Internet environment,which leads to unreliable geolocation results.To improve the accuracy and reliability of locating IP in real Internet,a street-level IP geolocation algorithm based on landmarks clustering is proposed.Firstly,we use the probes to measure the known landmarks to obtain their delay vectors,and cluster landmarks using them.Secondly,the landmarks are clustered again by their latitude and longitude,and the intersection of these two clustering results is taken to form training sets.Thirdly,we train multiple neural networks to get the mapping relationship between delay and location in each training set.Finally,we determine one of the neural networks for the target by the delay similarity and relative hop counts,and then geolocate the target by this network.As it brings together the delay and geographical coordinates clustering,the proposed algorithm largely improves the inconsistency between them and enhances the mapping relationship between them.We evaluate the algorithm by a series of experiments in Hong Kong,Shanghai,Zhengzhou and New York.The experimental results show that the proposed algorithm achieves street-level IP geolocation,and comparing with existing typical streetlevel geolocation algorithms,the proposed algorithm improves the geolocation reliability significantly.展开更多
基金supported by Metaverse Lab Program funded by the Ministry of Science and ICT(MSIT),and the Korea Radio Promotion Association(RAPA).
文摘The objective of style transfer is to maintain the content of an image while transferring the style of another image.However,conventional methods face challenges in preserving facial features,especially in Korean portraits where elements like the“Gat”(a traditional Korean hat)are prevalent.This paper proposes a deep learning network designed to perform style transfer that includes the“Gat”while preserving the identity of the face.Unlike traditional style transfer techniques,the proposed method aims to preserve the texture,attire,and the“Gat”in the style image by employing image sharpening and face landmark,with the GAN.The color,texture,and intensity were extracted differently based on the characteristics of each block and layer of the pre-trained VGG-16,and only the necessary elements during training were preserved using a facial landmark mask.The head area was presented using the eyebrow area to transfer the“Gat”.Furthermore,the identity of the face was retained,and style correlation was considered based on the Gram matrix.To evaluate performance,we introduced a metric using PSNR and SSIM,with an emphasis on median values through new weightings for style transfer in Korean portraits.Additionally,we have conducted a survey that evaluated the content,style,and naturalness of the transferred results,and based on the assessment,we can confidently conclude that our method to maintain the integrity of content surpasses the previous research.Our approach,enriched by landmarks preservation and diverse loss functions,including those related to“Gat”,outperformed previous researches in facial identity preservation.
基金supported by Princess Nourah bint Abdulrahman University Researchers Supporting Project number(PNURSP2022R196)Princess Nourah bint Abdulrahman University,Riyadh,Saudi Arabia.
文摘Imagery assessment is an efficient method for detecting craniofacial anomalies.A cephalometric landmark matching approach may help in orthodontic diagnosis,craniofacial growth assessment and treatment planning.Automatic landmark matching and anomalies detection helps face the manual labelling lim-itations and optimize preoperative planning of maxillofacial surgery.The aim of this study was to develop an accurate Cephalometric Landmark Matching method as well as an automatic system for anatomical anomalies classification.First,the Active Appearance Model(AAM)was used for the matching process.This pro-cess was achieved by the Ant Colony Optimization(ACO)algorithm enriched with proximity information.Then,the maxillofacial anomalies were classified using the Support Vector Machine(SVM).The experiments were conducted on X-ray cephalograms of 400 patients where the ground truth was produced by two experts.The frameworks achieved a landmark matching error(LE)of 0.50±1.04 and a successful landmark matching of 89.47%in the 2 mm and 3 mm range and of 100%in the 4 mm range.The classification of anomalies achieved an accuracy of 98.75%.Compared to previous work,the proposed approach is simpler and has a comparable range of acceptable matching cost and anomaly classification.Results have also shown that it outperformed the K-nearest neigh-bors(KNN)classifier.
文摘The facial landmarks can provide valuable information for expression-related tasks.However,most approaches only use landmarks for segmentation preprocessing or directly input them into the neural network for fully connection.Such simple combination not only fails to pass the spatial information to network,but also increases calculation amounts.The method proposed in this paper aims to integrate facial landmarks-driven representation into the triplet network.The spatial information provided by landmarks is introduced into the feature extraction process,so that the model can better capture the location relationship.In addition,coordinate information is also integrated into the triple loss calculation to further enhance similarity prediction.Specifically,for each image,the coordinates of 68 landmarks are detected,and then a region attention map based on these landmarks is generated.For the feature map output by the shallow convolutional layer,it will be multiplied with the attention map to correct the feature activation,so as to strengthen the key region and weaken the unimportant region.Finally,the optimized embedding output can be further used for downstream tasks.Three embeddings of three images output by the network can be regarded as a triplet representation for similarity computation.Through the CK+dataset,the effectiveness of such an optimized feature extraction is verified.After that,it is applied to facial expression similarity tasks.The results on the facial expression comparison(FEC)dataset show that the accuracy rate will be significantly improved after the landmark information is introduced.
文摘Australian Wool Innovation (AWI) today announced it had engaged world renowned style and innovation agency PeclersParis to provide colour and design trend direction for the Woolmark.Peclers’ team of fashion experts and trend forecasters this week
文摘New studies suggest that the Kaili Biota in southwestern China's Guizhou Province could be the third revealing evidence for the spectacular Cambrian Explosion of life about 530 million years ago, next to the Burgess Shale Biota of Canada, found in 1909,and the Chengjiang Biota, brought to light in Yunnan Province, China,in 1984.
基金supported by grants from National Natural Sciences Foundation of China (No. 30801314)China Hubei Provincial Science and Technology Department (No.2008CBD088)
文摘The accuracy and repeatability of computer aided cervical vertebra landmarking (CACVL) were investigated in cephalogram.120 adolescents (60 boys,60 girls) aged from 9.1 to 17.2 years old were randomly selected.Twenty-seven landmarks from the second to fifth cervical vertebrae on the lat-eral cephalogram were identified.In this study,the system of CACVL was developed and used to iden-tify and calculate the landmarks by fast marching method and parabolic curve fitting.The accuracy and repeatability in CACVL group were compared with those in two manual landmarking groups [orthodon-tic experts (OE) group and orthodontic novices (ON) group].The results showed that,as for the accu-racy,there was no significant difference between CACVL group and OE group no matter in x-axis or y-axis (P>0.05),but there was significant difference between CACVL group and ON group,as well as OE group and ON group in both axes (P<0.05).As for the repeatability,CACVL group was more reli-able than OE group and ON group in both axes.It is concluded that CACVL has the same or higher ac-curacy,better repeatability and less workload than manual landmarking methods.It’s reliable for cervi-cal parameters identification on the lateral cephalogram and cervical vertebral maturation prediction in orthodontic practice and research.
基金supported by the National Key Research and Development Program of China (No. 2019YFC1509701)the Second Tibetan Plateau Scientific Expedition and Research Program (STEP) (No. 2019QZKK0904)+1 种基金the National Natural Science Foundation of China (No. 42107184)the China Postdoctoral Science Foundation (No. 2018M640181)
文摘Despite extensive investigations,no precursor patterns for reliably predicting major earthquakes have thus far been identified.Seismogenic locked segments that can accumulate adequate strain energy to cause major earthquakes are highly heterogeneous and low brittle.The progressive cracking of the locked segments with these properties can produce an interesting seismic phenomenon:a landmark earthquake and a sequence of smaller subsequent earthquakes(pre-shocks)always arise prior to another landmark earthquake within a well-defined seismic zone and its current seismic period.Applying a mechanical model,magnitude constraint conditions,and case study data of 62 worldwide seismic zones,we show that two adjacent landmark earthquakes reliably occur at the volume-expansion point and peak-stress point(rupture)of a locked segment;thus,the former is an identified precursor for the latter.Such a precursor seismicity pattern before the locked-segment rupture has definite physical meanings,and it is universal regardless of the focal depth.Because the evolution of landmark earthquakes follows a deterministic rule described by the model,they are predictable.The results of this study lay a firm physical foundation for reliably predicting the occurrence of future landmark earthquakes in a seismic zone and can greatly improve our understanding of earthquake generation mechanism.
基金The work presented in this paper is supported by the National Key R&D Program of China[Nos.2016YFB0801303,2016QY01W0105]the National Natural Science Foundation of China[Nos.U1636219,U1804263,61602508,61772549,U1736214,61572052]Plan for Scientific Innovation Talent of Henan Province[No.2018JR0018].
文摘High-density street-level reliable landmarks are one of the important foundations for street-level geolocation.However,the existing methods cannot obtain enough street-level landmarks in a short period of time.In this paper,a street-level landmarks acquisition method based on SVM(Support Vector Machine)classifiers is proposed.Firstly,the port detection results of IPs with known services are vectorized,and the vectorization results are used as an input of the SVM training.Then,the kernel function and penalty factor are adjusted for SVM classifiers training,and the optimal SVM classifiers are obtained.After that,the classifier sequence is constructed,and the IPs with unknown service are classified using the sequence.Finally,according to the domain name corresponding to the IP,the relationship between the classified server IP and organization name is established.The experimental results in Guangzhou and Wuhan city in China show that the proposed method can be as a supplement to existing typical methods since the number of obtained street-level landmarks is increased substantially,and the median geolocation error using evaluated landmarks is reduced by about 2 km.
基金(in part)a grant in aid from the Emura Foundation for the Promotion of Cancer Research,No.01221
文摘AIM To characterize esophageal endoluminal landmarks to permit radial and longitudinal esophageal orientation and accurate lesion location.METHODS Distance from the incisors and radial orientation were estimated for the main left bronchus and the left atrium landmarks in 207 consecutive patients using white light examination. A sub-study was also performed using white light followed by endoscopic ultrasound(EUS) in 25 consecutive patients to confirm the findings.The scope orientation throughout the exam was maintained at the natural axis,where the left esophageal quadrant corresponds to the area between 6 and 9 o'clock. When an anatomical landmark was identified, it was recorded with a photograph and its quadrant orientation and distance from the incisors were determined. The reference points to obtain the distances and radial orientation were as follows: the midpoint of the left main bronchus and the most intense pulsatile zone of the left atrium. With the video processor system set to moderate insufflation, measurements were obtained at the end of the patients' air expiration.RESULTS The left main bronchus and left atrium esophageal landmarks were identified using white light in 99% and 100% of subjects at a mean distance of 25.8 cm(SD2.3), and 31.4 cm(SD 2.4) from the incisors, respectively. The left main bronchus landmark was found to be a tubular, concave, non-pulsatile, esophageal external compression, occupying approximately 1/4 of the circumference. The left atrium landmark was identified as a round, convex, pulsatile, esophageal external compression, occupying approximately 1/4 of the circumference. Both landmarks were identified using white light on the anterior esophageal quadrant. In the substudy, the left main bronchus was identified in 24(92%) patients at 25.4 cm(SD2.1) and 26.7 cm(SD 1.9) from the incisors, by white light and EUS, respectively.The left atrium was recognized in all patients at 30.5 cm(SD 1.9), and 31.6 cm(SD2.3) from the incisors, by both white light and EUS, respectively. EUS confirmed that the landmarks corresponded to these two structures, respectively, and that they were located on the anterior esophageal wall. The Bland-Altman plot demonstrated high agreement between the white light and EUS measurements.CONCLUSION This study provides an endoscopic characterization of esophageal landmarks corresponding to the left main bronchus and left atrium, to permit radial and longitudinal orientation and accurate lesion location.
基金This research was supported by National Natural Science Foundation of China(Nos.U1913603,61803251,51775322)National Key Research and Development Program of China(No.2019YFB1310003).
文摘The effectiveness of mobile robot aided for architectural construction depends strongly on its accurate localization ability.Localization of mobile robot is increasingly important for the printing of buildings in the construction scene.Although many available studies on the localization have been conducted,only a few studies have addressed the more challenging problem of localization for mobile robot in large-scale ongoing and featureless scenes.To realize the accurate localization of mobile robot in designated stations,we build an artificial landmark map and propose a novel nonlinear optimization algorithm based on graphs to reduce the uncertainty of the whole map.Then,the performances of localization for mobile robot based on the original and optimized map are compared and evaluated.Finally,experimental results show that the average absolute localization errors that adopted the proposed algorithm is reduced by about 21%compared to that of the original map.
基金the National Key R&D Program of China 2016YFB0801303(F.L.received the grant,the sponsors’website is https://service.most.gov.cn/)by the National Key R&D Program of China 2016QY01W0105(X.L.received the grant,the sponsors’website is https://service.most.gov.cn/)+5 种基金by the National Natural Science Foundation of China U1636219(X.L.received the grant,the sponsors’website is http://www.nsfc.gov.cn/)by the National Natural Science Foundation of China 61602508(J.L.received the grant,the sponsors’website is http://www.nsfc.gov.cn/)by the National Natural Science Foundation of China 61772549(F.L.received the grant,the sponsors’website is http://www.nsfc.gov.cn/)by the National Natural Science Foundation of China U1736214(F.L.received the grant,the sponsors’website is http://www.nsfc.gov.cn/)by the National Natural Science Foundation of China U1804263(X.L.received the grant,the sponsors’website is http://www.nsfc.gov.cn/)by the Science and Technology Innovation Talent Project of Henan Province 184200510018(X.L.received the grant,the sponsors’website is http://www.hnkjt.gov.cn/).
文摘Existing IP geolocation algorithms based on delay similarity often rely on the principle that geographically adjacent IPs have similar delays.However,this principle is often invalid in real Internet environment,which leads to unreliable geolocation results.To improve the accuracy and reliability of locating IP in real Internet,a street-level IP geolocation algorithm based on landmarks clustering is proposed.Firstly,we use the probes to measure the known landmarks to obtain their delay vectors,and cluster landmarks using them.Secondly,the landmarks are clustered again by their latitude and longitude,and the intersection of these two clustering results is taken to form training sets.Thirdly,we train multiple neural networks to get the mapping relationship between delay and location in each training set.Finally,we determine one of the neural networks for the target by the delay similarity and relative hop counts,and then geolocate the target by this network.As it brings together the delay and geographical coordinates clustering,the proposed algorithm largely improves the inconsistency between them and enhances the mapping relationship between them.We evaluate the algorithm by a series of experiments in Hong Kong,Shanghai,Zhengzhou and New York.The experimental results show that the proposed algorithm achieves street-level IP geolocation,and comparing with existing typical streetlevel geolocation algorithms,the proposed algorithm improves the geolocation reliability significantly.