Image Enhancement Of Elbow Vein And Design Of Puncture Vein Segment Visualization System

At present,venous puncture is one of the most basic nursing techniques in clinic.It is essential to find the location of venous puncture quickly and accurately.It is not easy for the medical workers who have little medical experience to find the puncture position of the patients whose veins are not obvious quickly and accurately,and the disputes between doctors and patients caused by it are also common.The vein of the human elbow is a common vein for daily medical blood collection and injection.This article uses the elbow vein image as a research object to develop a visualization system to assist venipuncture to help medical staff see it during the venipuncture Clearer,more accurate grasp of the puncture position.The main research contents are as follows:1.Construct an elbow vein image acquisition device,select the appropriate light source,camera and related accessories for the acquisition of vein image according to the characteristics of vein imaging.Collected 200 elbow vein images to form an image database for algorithm verification.2.Improve the problem of uneven illumination of vein images and much imaging noise through adaptive gamma transformation and bilateral filtering.The contrast limited adaptive histogram equalization(CLAHE)was used to initially improve the contrast of the elbow vein image,and a combination of multiple interpolation methods was used to eliminate the checkerboarding of the image after CLAHE processing to increase the speed of the algorithm.3.Design a multi-scale adaptive vascular filter based on Hessian matrix to filter the vein structure.Added adaptive parameter adjustment and contrast enhancement design to the filter.The experimental results show that the filter has a high extraction accuracy for vein structure.The superimposed fusion of the output image of the filter and the original image further improves the effect of vein observation.4.By segmenting the vein,based on the contour of the vein segment,starting from the shape,position,and gray value of the vein segment,extract 8-dimensional feature vectors for each vein segment.Through cross-validation,determine the classification model used to identify the puncture vein segment.The experimental results show that the classifier after integrating multiple classification models can achieve 91.7%accuracy.To realise the elbow vein image enhancement and the puncture vein segment visualization system.The system has no contact,clear venipuncture and simple operation.