Study On Section Analysis Based Automatic Nerve Fiber Diameter Calculation Method

As a basic unit of the nervous system's structure and function,neurons have always been the important research object of brain science.As the function is related to structure,the morphology of the neurons affects its function.Accurately obtaining the neuronal morphology is of great significance for the study of neuron classification and brain inspired simulation.A series of high-resolution optical imaging systems,such as Micro-Optical Sectioning Tomography,provide accurate and reliable data for observing the morphology of neurons.Based on the brain-wide high-resolution data of the mouse brain,we can get the neuron skeleton data by manual reconstruction.However,only skeleton data can't meet people's needs for neuronal morphology research,nerve fiber diameter is also a particularly important part of that.Accurate diameter calculation is of great significance for the study of neuronal morphological.Most of the existing diameter calculation methods make assumptions about the skeleton data and neuronal morphology.These assumptions often do not match the actual situation.,and making it difficult for these methods to achieve the desired results in real data.In addition,the skeleton data,which obtained by manual reconstruction,has the problems that the distance between the skeleton points is not uniform,and the skeleton trajectory is inferior to the neuron signal.These problems make it difficult to estimate the diameter of neuron and other related calculation processes.Aiming at the above problem,we propose a section analysis based automatic nerve fiber diameter calculation method,which is divided into two parts: neuronal skeleton correction and diameter calculation.The skeleton correction part includes some operations,such as skeleton point expansion,angle & signal strength joint constraint correction.Among them,the skeleton point expansion operation can solve the problem that the distance between the skeleton points is uneven;axial coordinate correction,coronal plane correction and angle & signal strength joint constraint correction can effectively improve the fitting degree of the skeleton data to the neuronal signal,and make the skeleton trajectory smoother.The diameter calculation part uses the cross-sectional information to approximate the diameter information of neuron.First,the neuron axis direction is obtained according to the position of the skeleton points,thereby obtaining the cross-sectional image.Then,the image is analyzed by threshold segmentation and connected component analysis to obtain the cross-sectional information of the neuron.Finally,the diameter is obtained approximately.The accuracy of the method was verified by testing on the real data.Based on the automatic calculation method,we developed a calculation tool,which make this method can be effectively applied to the neuron data with complex structures,such as the long-distance projection neurons.Then,the tool is accelerated by methods such as multi-thread parallel computing,which improves the calculation efficiency.Overall,in this paper,we propose a section analysis based automatic nerve fiber diameter calculation method.Using this method,we can obtain the diameter information of neurons effectively.