Accurate and continuous identification of individual cattle is crucial to precision farming in recent years.It is also the prerequisite to monitor the individual feed intake and feeding time of beef cattle at medium t...Accurate and continuous identification of individual cattle is crucial to precision farming in recent years.It is also the prerequisite to monitor the individual feed intake and feeding time of beef cattle at medium to long distances over different cameras.However,beef cattle can tend to frequently move and change their feeding position during feeding.Furthermore,the great variations in their head direction and complex environments(light,occlusion,and background)can also lead to some difficulties in the recognition,particularly for the bio-similarities among individual cattle.Among them,AlignedReID++model is characterized by both global and local information for image matching.In particular,the dynamically matching local information(DMLI)algorithm has been introduced into the local branch to automatically align the horizontal local information.In this research,the AlignedReID++model was utilized and improved to achieve the better performance in cattle re-identification(ReID).Initially,triplet attention(TA)modules were integrated into the BottleNecks of ResNet50 Backbone.The feature extraction was then enhanced through cross-dimensional interactions with the minimal computational overhead.Since the TA modules in AlignedReID++baseline model increased the model size and floating point operations(FLOPs)by 0.005 M and 0.05 G,the rank-1 accuracy and mean average precision(mAP)were improved by 1.0 percentage points and 2.94 percentage points,respectively.Specifically,the rank-1 accuracies were outperformed by 0.86 percentage points and 0.12 percentage points,respectively,compared with the convolution block attention module(CBAM)and efficient channel attention(ECA)modules,although 0.94 percentage points were lower than that of squeeze-and-excitation(SE)modules.The mAP metric values were exceeded by 0.22,0.86 and 0.12 percentage points,respectively,compared with the SE,CBAM,and ECA modules.Additionally,the Cross-Entropy Loss function was replaced with the CosFace Loss function in the global branch of baseline model.CosFace Loss and Hard Triplet Loss were jointly employed to train the baseline model for the better identification on the similar individuals.AlignedReID++with CosFace Loss was outperformed the baseline model by 0.24 and 0.92 percentage points in the rank-1 accuracy and mAP,respectively,whereas,AlignedReID++with ArcFace Loss was exceeded by 0.36 and 0.56 percentage points,respectively.The improved model with the TA modules and CosFace Loss was achieved in a rank-1 accuracy of 94.42%,rank-5 accuracy of 98.78%,rank-10 accuracy of 99.34%,mAP of 63.90%,FLOPs of 5.45 G,frames per second(FPS)of 5.64,and model size of 23.78 M.The rank-1 accuracies were exceeded by 1.84,4.72,0.76 and 5.36 percentage points,respectively,compared with the baseline model,part-based convolutional baseline(PCB),multiple granularity network(MGN),and relation-aware global attention(RGA),while the mAP metrics were surpassed 6.42,5.86,4.30 and 7.38 percentage points,respectively.Meanwhile,the rank-1 accuracy was 0.98 percentage points lower than TransReID,but the mAP metric was exceeded by 3.90 percentage points.Moreover,the FLOPs of improved model were only 0.05 G larger than that of baseline model,while smaller than those of PCB,MGN,RGA,and TransReID by 0.68,6.51,25.4,and 16.55 G,respectively.The model size of improved model was 23.78 M,which was smaller than those of the baseline model,PCB,MGN,RGA,and TransReID by 0.03,2.33,45.06,14.53 and 62.85 M,respectively.The inference speed of improved model on a CPU was lower than those of PCB,MGN,and baseline model,but higher than TransReID and RGA.The t-SNE feature embedding visualization demonstrated that the global and local features were achieve in the better intra-class compactness and inter-class variability.Therefore,the improved model can be expected to effectively re-identify the beef cattle in natural environments of breeding farm,in order to monitor the individual feed intake and feeding time.展开更多
Defect detection is vital in the nonwoven material industry,ensuring surface quality before producing finished products.Recently,deep learning and computer vision advancements have revolutionized defect detection,maki...Defect detection is vital in the nonwoven material industry,ensuring surface quality before producing finished products.Recently,deep learning and computer vision advancements have revolutionized defect detection,making it a widely adopted approach in various industrial fields.This paper mainly studied the defect detection method for nonwoven materials based on the improved Nano Det-Plus model.Using the constructed samples of defects in nonwoven materials as the research objects,transfer learning experiments were conducted based on the Nano DetPlus object detection framework.Within this framework,the Backbone,path aggregation feature pyramid network(PAFPN)and Head network models were compared and trained through a process of freezing,with the ultimate aim of bolstering the model's feature extraction abilities and elevating detection accuracy.The half-precision quantization method was used to optimize the model after transfer learning experiments,reducing model weights and computational complexity to improve the detection speed.Performance comparisons were conducted between the improved model and the original Nano Det-Plus model,YOLO,SSD and other common industrial defect detection algorithms,validating that the improved methods based on transfer learning and semi-precision quantization enabled the model to meet the practical requirements of industrial production.展开更多
The development of intelligent control techniques provides powerful means for the control of machine tools. In this paper, a intelligent control technique and an algorithm for precision control of CNC grinding of cera...The development of intelligent control techniques provides powerful means for the control of machine tools. In this paper, a intelligent control technique and an algorithm for precision control of CNC grinding of ceramic chips are introduced. In the process of ceramic chip CNC grinding, the dimension of the chips tends to get larger and the dimensional error to exceed the tolerance as the number of the chips increases which are machined on the same part program. There are many factors leading to the occurrence of the error and the law of error variation is very complicated. With the introduced intelligent self learning error compensation technique, the CNC system can improve the control strategy to compensate the error automatically. The simulational result is also given.展开更多
To prevent possible accidents,the study of data-driven analytics to predict hidden dangers in cloud service-based intelligent industrial production management has been the subject of increasing interest recently.A mac...To prevent possible accidents,the study of data-driven analytics to predict hidden dangers in cloud service-based intelligent industrial production management has been the subject of increasing interest recently.A machine learning algorithm that uses timeliness managing extreme learning machine is utilized in this article to achieve the above prediction.Compared with traditional learning algorithms,extreme learning machine(ELM) exhibits high performance because of its unique feature of a high generalization capability at a fast learning speed.Timeliness managing ELM is proposed by incorporating timeliness management scheme into ELM.When using the timeliness managing ELM scheme to predict hidden dangers,newly incremental data could be added prior to the historical data to maximize the contribution of the newly incremental training data,because the incremental data may be able to contribute reasonable weights to represent the current production situation according to practical analysis of accidents in some industrial productions.Experimental results from a coal mine show that the use of timeliness managing ELM can improve the prediction accuracy of hidden dangers with better stability compared with other similar machine learning methods.展开更多
Diagnosis and treatment of breast cancer have been improved during the last decade; however, breast cancer is still a leading cause of death among women in the whole world. Early detection and accurate diagnosis of th...Diagnosis and treatment of breast cancer have been improved during the last decade; however, breast cancer is still a leading cause of death among women in the whole world. Early detection and accurate diagnosis of this disease has been demonstrated an approach to long survival of the patients. As an attempt to develop a reliable diagnosing method for breast cancer, we integrated support vector machine (SVM), k-nearest neighbor and probabilistic neural network into a complex machine learning approach to detect malignant breast tumour through a set of indicators consisting of age and ten cellular features of fine-needle aspiration of breast which were ranked according to signal-to-noise ratio to identify determinants distinguishing benign breast tumours from malignant ones. The method turned out to significantly improve the diagnosis, with a sensitivity of 94.04%, a specificity of 97.37%, and an overall accuracy up to 96.24% when SVM was adopted with the sigmoid kernel function under 5-fold cross validation. The results suggest that SVM is a promising methodology to be further developed into a practical adjunct implement to help discerning benign and malignant breast tumours and thus reduce the incidence of misdiagnosis.展开更多
Electromagnetic Radiation Source Identification(ERSI) is a key technology that is widely used in military and radiation management and in electromagnetic interference diagnostics.The discriminative capability of machi...Electromagnetic Radiation Source Identification(ERSI) is a key technology that is widely used in military and radiation management and in electromagnetic interference diagnostics.The discriminative capability of machine learning methods has recently been used for facilitating ERSI.This paper presents a new approach to improve ERSI by adopting support vector machines,which are proven to be effective tools in pattern classification and regression,on the basis of the spatial distribution of electromagnetic radiation sources.Spatial information is converted from 3D cubes to 1D vectors with subscripts as inputs in order to simplify the model.The model is trained with 187 500 data sets in order to enable it to identify the types of radiation source types with an accuracy of up to 99.9%.The influence of parameters(e.g.,penalty parameter,reflection and noise from the ambient environment,and the scaling method for the input data) are discussed.The proposed method has good performance in noisy and reverberant environment.It has an identification accuracy of 82.15% when the signal-to-noise ratio is 20 dB.The proposed method has better accuracy in a noisy environment than artificial neural networks.Given that each Electromagnetic(EM) source has unique spatial characteristics,this method can be used for EM source identification and EM interference diagnostics.展开更多
Firstly, the early warning index system of coal mine safety production was given from four aspects as per- sonnel, environment, equipment and management. Then, improvement measures which are additional momentum method...Firstly, the early warning index system of coal mine safety production was given from four aspects as per- sonnel, environment, equipment and management. Then, improvement measures which are additional momentum method, adaptive learning rate, particle swarm optimization algorithm, variable weight method and asynchronous learning factor, are used to optimize BP neural network models. Further, the models are applied to a comparative study on coal mine safety warning instance. Results show that the identification precision of MPSO-BP network model is higher than GBP and PSO-BP model, and MPSO- BP model can not only effectively reduce the possibility of the network falling into a local minimum point, but also has fast convergence and high precision, which will provide the scientific basis for the forewarnin~ management of coal mine safetv production.展开更多
There are two fundamental goals in statistical learning： identifying relevant predictors and ensuring high prediction accuracy. The first goal, by means of variable selection, is of particular importance when the tru...There are two fundamental goals in statistical learning： identifying relevant predictors and ensuring high prediction accuracy. The first goal, by means of variable selection, is of particular importance when the true underlying model has a sparse representation. Discovering relevant predictors can enhance the performance of the prediction for the fitted model. Usually an estimate is considered desirable if it is consistent in terms of both coefficient estimate and variable selection. Hence, before we try to estimate the regression coefficients β , it is preferable that we have a set of useful predictors m hand. The emphasis of our task in this paper is to propose a method, in the aim of identifying relevant predictors to ensure screening consistency in variable selection. The primary interest is on Orthogonal Matching Pursuit（OMP）.展开更多
基金National Key Research and Development Program(2023YFD1301801)National Natural Science Foundation of China(32272931)+1 种基金Shaanxi Province Agricultural Key Core Technology Project(2024NYGG005)Shaanxi Province Key R&D Program(2024NC-ZDCYL-05-12)。
文摘Accurate and continuous identification of individual cattle is crucial to precision farming in recent years.It is also the prerequisite to monitor the individual feed intake and feeding time of beef cattle at medium to long distances over different cameras.However,beef cattle can tend to frequently move and change their feeding position during feeding.Furthermore,the great variations in their head direction and complex environments(light,occlusion,and background)can also lead to some difficulties in the recognition,particularly for the bio-similarities among individual cattle.Among them,AlignedReID++model is characterized by both global and local information for image matching.In particular,the dynamically matching local information(DMLI)algorithm has been introduced into the local branch to automatically align the horizontal local information.In this research,the AlignedReID++model was utilized and improved to achieve the better performance in cattle re-identification(ReID).Initially,triplet attention(TA)modules were integrated into the BottleNecks of ResNet50 Backbone.The feature extraction was then enhanced through cross-dimensional interactions with the minimal computational overhead.Since the TA modules in AlignedReID++baseline model increased the model size and floating point operations(FLOPs)by 0.005 M and 0.05 G,the rank-1 accuracy and mean average precision(mAP)were improved by 1.0 percentage points and 2.94 percentage points,respectively.Specifically,the rank-1 accuracies were outperformed by 0.86 percentage points and 0.12 percentage points,respectively,compared with the convolution block attention module(CBAM)and efficient channel attention(ECA)modules,although 0.94 percentage points were lower than that of squeeze-and-excitation(SE)modules.The mAP metric values were exceeded by 0.22,0.86 and 0.12 percentage points,respectively,compared with the SE,CBAM,and ECA modules.Additionally,the Cross-Entropy Loss function was replaced with the CosFace Loss function in the global branch of baseline model.CosFace Loss and Hard Triplet Loss were jointly employed to train the baseline model for the better identification on the similar individuals.AlignedReID++with CosFace Loss was outperformed the baseline model by 0.24 and 0.92 percentage points in the rank-1 accuracy and mAP,respectively,whereas,AlignedReID++with ArcFace Loss was exceeded by 0.36 and 0.56 percentage points,respectively.The improved model with the TA modules and CosFace Loss was achieved in a rank-1 accuracy of 94.42%,rank-5 accuracy of 98.78%,rank-10 accuracy of 99.34%,mAP of 63.90%,FLOPs of 5.45 G,frames per second(FPS)of 5.64,and model size of 23.78 M.The rank-1 accuracies were exceeded by 1.84,4.72,0.76 and 5.36 percentage points,respectively,compared with the baseline model,part-based convolutional baseline(PCB),multiple granularity network(MGN),and relation-aware global attention(RGA),while the mAP metrics were surpassed 6.42,5.86,4.30 and 7.38 percentage points,respectively.Meanwhile,the rank-1 accuracy was 0.98 percentage points lower than TransReID,but the mAP metric was exceeded by 3.90 percentage points.Moreover,the FLOPs of improved model were only 0.05 G larger than that of baseline model,while smaller than those of PCB,MGN,RGA,and TransReID by 0.68,6.51,25.4,and 16.55 G,respectively.The model size of improved model was 23.78 M,which was smaller than those of the baseline model,PCB,MGN,RGA,and TransReID by 0.03,2.33,45.06,14.53 and 62.85 M,respectively.The inference speed of improved model on a CPU was lower than those of PCB,MGN,and baseline model,but higher than TransReID and RGA.The t-SNE feature embedding visualization demonstrated that the global and local features were achieve in the better intra-class compactness and inter-class variability.Therefore,the improved model can be expected to effectively re-identify the beef cattle in natural environments of breeding farm,in order to monitor the individual feed intake and feeding time.
基金National Key Research and Development Program of China(Nos.2022YFB4700600 and 2022YFB4700605)National Natural Science Foundation of China(Nos.61771123 and 62171116)+1 种基金Fundamental Research Funds for the Central UniversitiesGraduate Student Innovation Fund of Donghua University,China(No.CUSF-DH-D-2022044)。
文摘Defect detection is vital in the nonwoven material industry,ensuring surface quality before producing finished products.Recently,deep learning and computer vision advancements have revolutionized defect detection,making it a widely adopted approach in various industrial fields.This paper mainly studied the defect detection method for nonwoven materials based on the improved Nano Det-Plus model.Using the constructed samples of defects in nonwoven materials as the research objects,transfer learning experiments were conducted based on the Nano DetPlus object detection framework.Within this framework,the Backbone,path aggregation feature pyramid network(PAFPN)and Head network models were compared and trained through a process of freezing,with the ultimate aim of bolstering the model's feature extraction abilities and elevating detection accuracy.The half-precision quantization method was used to optimize the model after transfer learning experiments,reducing model weights and computational complexity to improve the detection speed.Performance comparisons were conducted between the improved model and the original Nano Det-Plus model,YOLO,SSD and other common industrial defect detection algorithms,validating that the improved methods based on transfer learning and semi-precision quantization enabled the model to meet the practical requirements of industrial production.
文摘The development of intelligent control techniques provides powerful means for the control of machine tools. In this paper, a intelligent control technique and an algorithm for precision control of CNC grinding of ceramic chips are introduced. In the process of ceramic chip CNC grinding, the dimension of the chips tends to get larger and the dimensional error to exceed the tolerance as the number of the chips increases which are machined on the same part program. There are many factors leading to the occurrence of the error and the law of error variation is very complicated. With the introduced intelligent self learning error compensation technique, the CNC system can improve the control strategy to compensate the error automatically. The simulational result is also given.
基金partially supported by the National Key Technologies R&D Program of China under Grant No.2015BAK38B01the National Natural Science Foundation of China under Grant Nos.61174103 and 61272357the Fundamental Research Funds for the Central Universities under Grant No.06500025
文摘To prevent possible accidents,the study of data-driven analytics to predict hidden dangers in cloud service-based intelligent industrial production management has been the subject of increasing interest recently.A machine learning algorithm that uses timeliness managing extreme learning machine is utilized in this article to achieve the above prediction.Compared with traditional learning algorithms,extreme learning machine(ELM) exhibits high performance because of its unique feature of a high generalization capability at a fast learning speed.Timeliness managing ELM is proposed by incorporating timeliness management scheme into ELM.When using the timeliness managing ELM scheme to predict hidden dangers,newly incremental data could be added prior to the historical data to maximize the contribution of the newly incremental training data,because the incremental data may be able to contribute reasonable weights to represent the current production situation according to practical analysis of accidents in some industrial productions.Experimental results from a coal mine show that the use of timeliness managing ELM can improve the prediction accuracy of hidden dangers with better stability compared with other similar machine learning methods.
基金Joint Research Project Between Chongqing University and National University of Singapore (No. ARF-151-000-014-112)the Basic Research & Applied Basic Research Program of Chongqing University (No.71341103)Natural Science Foundation of Chongqing S & T Committee(No. CSTC,2006BB5240)
文摘Diagnosis and treatment of breast cancer have been improved during the last decade; however, breast cancer is still a leading cause of death among women in the whole world. Early detection and accurate diagnosis of this disease has been demonstrated an approach to long survival of the patients. As an attempt to develop a reliable diagnosing method for breast cancer, we integrated support vector machine (SVM), k-nearest neighbor and probabilistic neural network into a complex machine learning approach to detect malignant breast tumour through a set of indicators consisting of age and ten cellular features of fine-needle aspiration of breast which were ranked according to signal-to-noise ratio to identify determinants distinguishing benign breast tumours from malignant ones. The method turned out to significantly improve the diagnosis, with a sensitivity of 94.04%, a specificity of 97.37%, and an overall accuracy up to 96.24% when SVM was adopted with the sigmoid kernel function under 5-fold cross validation. The results suggest that SVM is a promising methodology to be further developed into a practical adjunct implement to help discerning benign and malignant breast tumours and thus reduce the incidence of misdiagnosis.
基金supported by the National Natural Science Foundation of China under Grant No.61201024
文摘Electromagnetic Radiation Source Identification(ERSI) is a key technology that is widely used in military and radiation management and in electromagnetic interference diagnostics.The discriminative capability of machine learning methods has recently been used for facilitating ERSI.This paper presents a new approach to improve ERSI by adopting support vector machines,which are proven to be effective tools in pattern classification and regression,on the basis of the spatial distribution of electromagnetic radiation sources.Spatial information is converted from 3D cubes to 1D vectors with subscripts as inputs in order to simplify the model.The model is trained with 187 500 data sets in order to enable it to identify the types of radiation source types with an accuracy of up to 99.9%.The influence of parameters(e.g.,penalty parameter,reflection and noise from the ambient environment,and the scaling method for the input data) are discussed.The proposed method has good performance in noisy and reverberant environment.It has an identification accuracy of 82.15% when the signal-to-noise ratio is 20 dB.The proposed method has better accuracy in a noisy environment than artificial neural networks.Given that each Electromagnetic(EM) source has unique spatial characteristics,this method can be used for EM source identification and EM interference diagnostics.
文摘Firstly, the early warning index system of coal mine safety production was given from four aspects as per- sonnel, environment, equipment and management. Then, improvement measures which are additional momentum method, adaptive learning rate, particle swarm optimization algorithm, variable weight method and asynchronous learning factor, are used to optimize BP neural network models. Further, the models are applied to a comparative study on coal mine safety warning instance. Results show that the identification precision of MPSO-BP network model is higher than GBP and PSO-BP model, and MPSO- BP model can not only effectively reduce the possibility of the network falling into a local minimum point, but also has fast convergence and high precision, which will provide the scientific basis for the forewarnin~ management of coal mine safetv production.
文摘There are two fundamental goals in statistical learning： identifying relevant predictors and ensuring high prediction accuracy. The first goal, by means of variable selection, is of particular importance when the true underlying model has a sparse representation. Discovering relevant predictors can enhance the performance of the prediction for the fitted model. Usually an estimate is considered desirable if it is consistent in terms of both coefficient estimate and variable selection. Hence, before we try to estimate the regression coefficients β , it is preferable that we have a set of useful predictors m hand. The emphasis of our task in this paper is to propose a method, in the aim of identifying relevant predictors to ensure screening consistency in variable selection. The primary interest is on Orthogonal Matching Pursuit（OMP）.
