3D Spatial Localization Of EEG Electrodes

Electroencephalogram(EEG)is an important method in the diagnosis of mental diseases.It plays an irreplaceable role in clinical diagnosis of shock,epilepsy and other diseases.In the process of EEG diagnosis,an important step is to locate the location of the brain power signal.Locating the spatial position of EEG electrodes has become a research subject of great significance.In recent years,many types of solutions and experimental studies have been designed and implemented to accurately locate the spatial location of EEG electrodes.All methods are hard to weigh in accuracy,convenience,stability,and no electrical signal interference.Photogrammetry has become the most popular research method because it can work without interference from electrical signals.However,it can not be fully satisfied in cost,simplicity and accuracy.In this paper,a new photogrammetric method based on depth camera is proposed by comparing photogrammetry methods.The main work are as follows:1.A three-dimensional positioning system of EEG electrodes based on RGB-D multi-mode data is constructed in this paper.A combination of a traditional CCD industrial camera and a new depth camera TOF are used to obtain the color and depth information in the scene.The system shoots at 5 angles so that all electrode positions can be captured by both cameras units.2.In this paper,we designed a calibration plate for depth camera depth features and proposed a new calibration method for depth camera and color camera.Due to the depth map acquired by the depth camera has the characteristics of low resolution,voids and so on,the calibration method using the traditional checkerboard calibration method has larger error.In this paper,a 2.5D calibration board with depth information is designed to obtain more accurate depth information.In the calibration process,the system uses accurate 3D point clouds instead of distorted depth amplitude images,which greatly improves the calibration accuracy.3.The electrode cap recognition and extraction,point cloud stitching and brain 3D reconstruction algorithms are realized in this paper.An automatic detection algorithm based on connected area threshold is used to recognize and extract electrodes,which can identify the poles and remove other interference points which do not meet the requirements.For electrode points and point cloud,the SVD algorithm is used instead of the ICP algorithm which is not suitable for large angle point cloud conversion to calculate the conversion relationship.4.On the basis of the head model experiment,this paper adds the real person contrast experiments and the contrast experiments of different types of depth camera.Comparative experiments provide more reliable data for the system and test system accuracy.In contrast experiment,in addition to the contrast data before and after the camera calibration,two real people experiments data are added to increase the reliability of the data.In this paper,a new depth camera with higher resolution is added to improve the accuracy of the system.