Research On Key Techniques Of Optical Projection Tomography For In-Vivo Imaging

Optical 3D imaging technology is widely used because of its advantages of no radiation,high resolution and strong contrast.It is apllied to life science,medical pharmacy and other research fields.Optical projection Tomography(OPT)is an optically spectrum of X-ray computed tomography(CT)technique in which parallel ray projection is assumed and 3D images can be reconstructed using filtered back projection(FBP)algorithms.This technique is mainly applicable to the visual observation of mesoscopic biological systems(such as intact embryos,in vitro organs and living small model animals,etc.),and is playing an increasingly important role in the field of biomedical research.However,when it is applied to in-vivo imaging of biological tissues,the high scattering of biological tissues restricts the application of OPT technology.As a result,OPT technology is mainly applied to in-vitro imaging treated by chemical cleaning and transparency or biological embryos treated by transgenic,which greatly restricts the scope of application of OPT technology.In this paper,an OPT experimental device for visible spectrum based on OPT adaptor was designed.The design method and implementation method of the OPT experimental device were explored in the early stage,which laid a foundation for the construction of the OPT experimental system for living tissue in near infrared spectrum.Then the imaging depth was improved by reducing the light scattering of biological tissues during the imaging process.Then the OPT in vivo imaging speed was improved by reducing the number of projection images required for reconstruction.Through the post-image processing and enhancement algorithm,the effect of the reconstructed image is improved.Through the above research,the imaging effect of in vivo imaging is improved.The main research contents and contributions of the paper are as follows:1.OPT imaging systems require imaging depth to cover half of the sample.Traditional OPT imaging systems use a low numerical aperture objective for observation to obtain sufficient depth of field.In this paper,the OPT experimental system is developed and designed based on the OPT adapter.Firstly,the change rule of projection image when focal plane is located at different positions is analyzed theoretically.Then the two dimensional image sequences are observed in different focal planes and analyzed and compared.Finally,a microfluidic system is used to realize dynamic imaging of samples at specific flow rates.Through the design and implementation of OPT imaging system of visible spectrum,this part of work provides the experimental basis and algorithm support for the design and implementation of the OPT experimental device for imaging biological tissues in near infrared spectrum.2.As the problems caused by light scattering,OPT imaging penetration depth and imaging resolution are reduced.In this paper,short-wave infrared(SWIR)light is used as the light source instead of visible light to realize in-vivo OPT imaging of biological tissue in SWIR spectrum.Theoretical analysis shows that the NIR-Ⅱ window(1000-1700nm)has many inherent properties suitable for in-vivo imaging compared with the NIR-Ⅰ window(7501000nm),such as deeper imaging depth,lower auto fluorescence and less biological tissue scattering,which contribute to improved resolution and higher signal-to-noise ratio.The NIRⅡa window(1300-1400nm)and the NIR-Ⅱb window(1500-1700nm)have great potential for in-vivo biomedical imaging because these spectrums avoid strong water absorption.In this paper,the OPT imaging device independently developed and designed is used to obtain the OPT projection image sequence.The device can realize 360° high-precision rotation of the sample.After software and hardware design and optimization,the experimental prototype only needs 18s to obtain the OPT image sequence,greatly improving the imaging speed of the OPT device,which is of great significance for in-vivo imaging.Our work focuses on the near infrared NIR-Ⅱb window,which has wavelengths ranging from 1500 to 1700 nm.A 1550nm laser was used as the light source to obtain the projected images.In order to solve the problem of unmatched rotation centers.the center of rotation correction and FBP algorithm were used to complete the reconstruction.The results show that,compared with visible light,OPT imaging in 1550nm spectrum greatly improves the imaging depth,and the depth resolution of the acquired three-dimensional volume can reach 50 μm.The advantages of 1550 nm spectrum are mainly reflected in reduced scattering,lower spontaneous fluorescence and deeper penetration depth,thus achieving higher imaging contrast.Through the acquired projection images,the internal bone structure and arterial vessel information of zebrafish can be clearly seen in the short-wave infrared spectrum,which is of great significance for biological tissue imaging.3.The SIRT algorithm is applied to the reconstruction of OPT when the number of projections is sparse in this paper.By reducing the number of projections required for reconstruction,the image acquisition time is reduced and the imaging speed of OPT is improved.The FBP algorithm can get good results in the case of intensive OPT projection quantity,but reconstruction from under-sampled OPT data may lead to artifacts and image quality degradation.In order to remove artifacts and improve image quality,simultaneous iterative reconstruction technology(SIRT)was introduced in this paper.The image quality of FBP and SIRT were compared respectively,and the structure similarity(SSIM)and peak signal-to-noise ratio(PSNR)were calculated.By comparing the reconstruction results of SIRT algorithm and FBP algorithm on the simulation data and living zebrafish embryos,the experimental results show that SIRT is always superior to FBP in reducing artifacts and enhancing image contrast，especially in the case of a small number of projections.The SIRT method enables high quality reconstruction with 50 or fewer projections,thereby significantly reducing image acquisition time and light dose while maintaining reconstruction quality.Through optimization and GPU acceleration,the SIRT algorithm can converge faster,thus reducing image processing time.Experimental results show that the reconstruction result of SIRT algorithm is better than that of FBP algorithm when the number of projections is reduced.In addition,SIRT were better at maintaining the display of vascular information in the sample,which has important implications for monitoring angiogenesis.4.In order to improve the effect of in-vivo imaging,the method based on HDR and SIRT is used to enhance the image of OPT projection sequence acquired in short-wave infrared spectrum.The method based on HDR and SIRT is also used to improve the quality of reconstructed images.In this part,the method of HDR fusion is firstly used to fuse twodimensional projection images of different brightness,so as to generate HDR fusion projection images at each Angle,and finally generate HDR projection image sequence.Then the generated HDR projection image sequence is reconstructed,and finally the HDR-OPT 3D reconstruction is realized in SWIR spectrum.The reconstructed image of OPT based on HDR can simultaneously retain the details of the bright and dark parts of zebrafish,making the acquired projection image and reconstructed image clearer and showing more details.The contrast improvement of image enhancement was objectively analyzed through three-dimensional gray distribution and histogram statistics.More importantly,the HDR method can recover the loss of key information caused by overexposure,which is of great significance for biological tissue imaging.SIRT algorithm uses fewer number of projection images to achieve better reconstruction,which is specifically manifested as fewer artifacts,higher image contrast and less distortion in some areas.In summary,this thesis mainly studies the OPT in-vivo imaging technology from three aspects:improving the speed of in-vivo imaging.improving the depth of in-vivo imaging,and enhancing the effect of in-vivo imaging reconstruction.The main innovations and contributions of this paper are summarized as follows:1.The OPT projection image acquisition and reconstruction of near infrared spectrum based on InGaAs infrared camera.The OPT projection image sequence in the near-infrared spectrum was obtained by the self-developed OPT experimental device.The problem of axis mismatch in the OPT projection image sequence is solved by an adaptive center of rotation correction algorithm,and OPT reconstruction images in the 1550nm spectrum are obtained.The experimental results show that OPT imaging in the NIR spectrum improves the in-vivo OPT imaging depth.2.Implement OPT sparse projection image reconstruction based on SIRT iterative reconstruction algorithm.The SIRT iterative reconstruction algorithm can use 50 or less projection images to achieve OPT high-quality reconstruction,which can reduce artifacts and improve the contrast of the reconstructed image.This method reduces the number of projections required for reconstruction,thereby reducing the image acquisition time and finally improving the imaging speed.3.The reconstruction image enhance method based on HDR and SIRT algorithms were applied to OPT reconstruction in the near-infrared spectrum.In this paper,the reconstruction image of HDR-OPT in the 1550 nm spectrum is achieved based on the HDR method,and OPT image reconstruction in the 1550 nm spectrum is obtained based on the SIRT algorithm.The image artifacts are reduced and image contrast is enhanced.