Design And Research Of Medical Ultrasound Image Processing Simulation Platform Based On MATLAB

With the improvement of the quality of life and the development of modern medical technology,technology assists humans to achieve more and more powerful solving capabilities.Medical image inspection technology is an essential auxiliary technology in the medical field today.Medical ultrasound examination is the first choice for primary screening because it is non-invasive,has high time resolution,and is easy to implement at low cost.However,ultrasound images are different from CT and magnetic resonance images,and because of its imaging principle,they are more susceptible to external environmental factors and equipment.Parameters and other constraints affect the image,resulting in problems such as unclear images,low contrast,uneven grayscale,and blurred organizational boundaries.The image processing software that comes with medical equipment can only meet the requirements for doctors to simply process images on medical equipment.It does not provide an academic research platform.At present,most medical images are in DICOM(Digital Imaging Communications in Medicine)format.General image processing software cannot directly read and process.Therefore,the development of a simulation platform capable of reading and improving the quality of medical ultrasound images has become a practical need for lesion research.This paper designed a set of medical ultrasound image processing simulation platform based on MATLAB.The platform has 38 functions of 6 modules including DICOM format image enhancement,image filtering,image segmentation,morphological processing and edge detection.In addition,the Pulse Coupled Neural Network(PCNN)image segmentation is innovatively incorporated into the platform,and the maximum number of iterations is determined using the maximum rule of image information entropy.Finally,two types of typical and representative stone lesion images in the medical ultrasound images of human gallbladder stones were used as experimental images.The simulation platform was tested and evaluated by a combination of subjective evaluation and objectiveevaluation.It is suitable for the process of medical ultrasound image of human gallbladder stones.The experimental results show that:(1)Image enhancement module: After the two test images are processed by the image enhancement module in the platform,the gray level range of the image is adjusted,and the information of the stones that need to be observed is strengthened.The unimportant,The influential information was weakened,which solved the problems of low contrast and uneven grayscale of the ultrasound image.(2)Image filtering module: filter the two test images with four different algorithms in the platform,and set the window size to 5×5 uniformly,using the objective evaluation index signal to noise ratio(SNR)To verify the effectiveness of the four filtering effects,after testing,the signal-to-noise ratio of the original test image cholesterol stone ultrasound image is7.951 dB,and the signal-to-noise ratio after mean filtering,median filtering,Gaussian smoothing filtering and Wiener filtering are improved It is 12.496 dB,18.095 dB,23.152 dB,28.345 dB,indicating that the platform filter module can effectively improve the clarity of the ultrasound image.(3)Image segmentation module: use the three segmentation algorithms of threshold segmentation,region segmentation and PCNN segmentation to segment the stone parts of the two test images,respectively,and register the results with the doctor's manual segmentation image for registration.The results show that PCNN segmentation It has a recognition ability similar to that of the human brain,and it has the best coincidence with manual segmentation;the threshold segmentation effect is larger than that of PCNN segmentation,and the area of ??over-segmentation is larger,but the threshold segmentation algorithm has a small amount of calculation and is easier to implement;regional segmentation stone contour It is not smooth enough and clear,and there is over-segmentation inside the stone,but regional segmentation can be achieved by positioning segmentation.(4)Morphology processing module: use the expansion,erosion,open operation and close operation in this module to process the image after the area is divided,and the result is registered with the image manually divided by the doctor to register,and the area after the morphological processing can be found The segmentedimage fills the over-segmented parts,so that the region segmentation improves the segmentation accuracy while positioning the segmentation.(5)Edge detection module:perform the edge detection of Roberts operator,Sobel operator,Prewitt operator,Canny operator and LoG operator on the two test images,and use the edge detection operator proposed by Pratt to objectively evaluate the quality of quantified indicators The factor R is used as the objective evaluation standard,and the doctor manually divides the image as the subjective evaluation standard.The results show that the five operators can achieve accurate edge detection,and the subjective and objective evaluations are consistent.In summary,the processed image has better visibility,which effectively improves the image quality of the original image and enriches the amount of image information.This platform not only implements a variety of medical graphics processing methods for medical staff and researchers,expands the methods of medical image research,but also provides an effective reference for the research and improvement of related medical image processing technology algorithms;at the same time,for non-programming For professional medical workers,the platform has a friendly interface,flexible parameter adjustment,and simple operation.It can be converted into an.exe executable file and run out of the MATLAB environment.