As 3D acquisition technology develops and 3D sensors become increasingly affordable,large quantities of 3D point cloud data are emerging.How to effectively learn and extract the geometric features from these point clo...As 3D acquisition technology develops and 3D sensors become increasingly affordable,large quantities of 3D point cloud data are emerging.How to effectively learn and extract the geometric features from these point clouds has become an urgent problem to be solved.The point cloud geometric information is hidden in disordered,unstructured points,making point cloud analysis a very challenging problem.To address this problem,we propose a novel network framework,called Tree Graph Network(TGNet),which can sample,group,and aggregate local geometric features.Specifically,we construct a Tree Graph by explicit rules,which consists of curves extending in all directions in point cloud feature space,and then aggregate the features of the graph through a cross-attention mechanism.In this way,we incorporate more point cloud geometric structure information into the representation of local geometric features,which makes our network perform better.Our model performs well on several basic point clouds processing tasks such as classification,segmentation,and normal estimation,demonstrating the effectiveness and superiority of our network.Furthermore,we provide ablation experiments and visualizations to better understand our network.展开更多
Experimental and theoretical studies have reported that the precise firing of neurons is crucial for sensory representation.Autapse serves as a special synapse connecting neuron and itself,which has also been found to...Experimental and theoretical studies have reported that the precise firing of neurons is crucial for sensory representation.Autapse serves as a special synapse connecting neuron and itself,which has also been found to improve the accuracy of neuronal response.In current work,the effect of autaptic delay signal on the spike-timing precision is investigated on a single autaptic Hodgkin–Huxley neuron in the present of noise.The simulation results show that both excitatory and inhibitory autaptic signals can effectively adjust the precise spike time of neurons with noise by choosing the appropriate coupling strength g and time delay of autaptic signalτ.The g–τparameter space is divided into two regions:one is the region where the spike-timing precision is effectively regulated;the other is the region where the neuronal firing is almost not regulated.For the excitatory and inhibitory autapse,the range of parameters causing the accuracy of neuronal firing is different.Moreover,it is also found that the mechanisms of the spike-timing precision regulation are different for the two kinds of autaptic signals.展开更多
基金supported by the National Natural Science Foundation of China (Grant Nos.91948203,52075532).
文摘As 3D acquisition technology develops and 3D sensors become increasingly affordable,large quantities of 3D point cloud data are emerging.How to effectively learn and extract the geometric features from these point clouds has become an urgent problem to be solved.The point cloud geometric information is hidden in disordered,unstructured points,making point cloud analysis a very challenging problem.To address this problem,we propose a novel network framework,called Tree Graph Network(TGNet),which can sample,group,and aggregate local geometric features.Specifically,we construct a Tree Graph by explicit rules,which consists of curves extending in all directions in point cloud feature space,and then aggregate the features of the graph through a cross-attention mechanism.In this way,we incorporate more point cloud geometric structure information into the representation of local geometric features,which makes our network perform better.Our model performs well on several basic point clouds processing tasks such as classification,segmentation,and normal estimation,demonstrating the effectiveness and superiority of our network.Furthermore,we provide ablation experiments and visualizations to better understand our network.
基金the Fundamental Research Funds for the Central Universities,China(Grant No.GK201903020)the National Natural Science Foundation of China(Grant No.12005006)Scientific research project of Education Department of Gansu Province,China(Grant No.2016A-049).
文摘Experimental and theoretical studies have reported that the precise firing of neurons is crucial for sensory representation.Autapse serves as a special synapse connecting neuron and itself,which has also been found to improve the accuracy of neuronal response.In current work,the effect of autaptic delay signal on the spike-timing precision is investigated on a single autaptic Hodgkin–Huxley neuron in the present of noise.The simulation results show that both excitatory and inhibitory autaptic signals can effectively adjust the precise spike time of neurons with noise by choosing the appropriate coupling strength g and time delay of autaptic signalτ.The g–τparameter space is divided into two regions:one is the region where the spike-timing precision is effectively regulated;the other is the region where the neuronal firing is almost not regulated.For the excitatory and inhibitory autapse,the range of parameters causing the accuracy of neuronal firing is different.Moreover,it is also found that the mechanisms of the spike-timing precision regulation are different for the two kinds of autaptic signals.