Design And Implementation Of Forward/Backward Filtering Denoising Algorithm In Photoacoustic Image Process

Photoacoustic Imaging(PAI)is an emerging imaging technology that combines the advantages of pure optical imaging with high contrast and ultrasound imaging depth.Compared with other traditional biological imaging,PAI has its wavelength and frequency.It is often used for the imaging of cartilage,brain and muscle tissue parts,and can produce considerable imaging results for the target body parts with resolution and contrast performance.However,in the photoacoustic imaging process,due to environmental and equipment factors,there will be noise and clutter in the image,resulting in low final signal-to-noise ratio and low resolution.Therefore,how to eliminate potential noise interference during photoacoustic imaging is the key to improving the photoacoustic imaging effect.In this paper,different denoising algorithms are designed from the two aspects of the forward and backward processes of photoacoustic imaging to improve the photoacoustic signal and the final image quality.The specific research contents are as follows:(1)A photoacoustic image reconstruction algorithm based on Renyi entropy is proposed.The reconstruction filtering algorithm determines the threshold used for segmentation according to the distribution of the original photoacoustic signal Renyi entropy,and then filters out the existing clutter;then uses the photoacoustic signal filtered by the previous step to filter out the clutter Photoacoustic image reconstruction.The algorithm is used to process samples of different dimensions,including zero-dimensional pencil core cross section,one-dimensional hair silk and two-dimensional rat cerebral cortical blood vessels.The final results of the experiment show that before using the Renyi entropy filtering algorithm to shoot early Compared with the reconstructed photoacoustic images,the difference in brightness level and resolution are increased by 32.45% and 30.78%,respectively,and the mean square error is reduced by 35.01% on the original basis,and the signal-to-noise ratio is increased by 47.66%.(2)Design CS / SVM combination algorithm to improve the quality of photoacoustic images.The photoacoustic image is decomposed by wavelet transform,and the high frequency coefficient of the image is measured and restored by using the measurement matrix and the matching tracking algorithm.The reconstruction coefficient of the high frequency component and the original coefficient of the low frequency component are combined into a new wavelet coefficient.All wavelet coefficients are processed by the trained classification model.The classification wavelet coefficients are divided into two parts: noiseand non-noise,and the noise part is processed by the soft threshold.Finally,inverse wavelet reconstruction is used to obtain the denoised image.The results showed that the contrast of the vocal images of the cerebral cortex of mice increased by 39.19%,the signal-to-noise ratio increased by 59.71%,and the MSE decreased by 20.83%.At the same time,compared with the traditional medical image processing combination algorithm,the designed CS / SVM combination algorithm greatly improves the signal-to-noise ratio of the sample(at least 26.95%),further reducing the MSE of the image(about 11.98%).