激光散斑衬比血流成像关键技术及应用研究进展

翟林君; 傅玉青; 杜永兆

doi:doi:10.3788/CJL221200

中国激光, 2023, 50 (9): 0907106, 网络出版: 2023-03-06

激光散斑衬比血流成像关键技术及应用研究进展下载： 1232次

Advances in Laser Speckle Contrast Imaging: Key Techniques and Applications

翟林君 ¹傅玉青 ²杜永兆 ^1,2,*

作者单位

¹ 华侨大学生物医学学院，福建泉州 362021

² 华侨大学工学院，福建泉州 362021

图 & 表

图 1. LSCI实验系统示意图^［30］

Fig. 1. Schematic setup for laser speckle contrast imaging (LSCI)^[30]

下载图片查看原文

图 2. LSCI关键技术问题的分析及解决思路

Fig. 2. Analysis and solution of key technical problems of LSCI

下载图片查看原文

图 3. aLSCI算法模型^［98］

Fig. 3. Scheme of aLSCI algorithm^[98]

下载图片查看原文

图 4. 不同算法的实验结果对比^［98］。（a）tLSCI（时间衬比分析法）；（b）sLSCI（空间衬比分析法）；（c）stLSCI（时空联合衬比分析法）；（d）savgtLSCI（空间平均衬比分析法）；（e）tavgsLSCI（时间平均衬比分析法）；（f）aLSCI；（g）不同算法的对比度噪声比

Fig. 4. Comparative experimental results of different algorithms^[98]. (a) tLSCI algorithm; (b) sLSCI algorithm; (c) stLSCI algorithm; (d) savgtLSCI algorithm; (e) tavgsLSCI algorithm; (f) aLSCI algorithm; (g) contrast-to-noise ratio (CNR) of different algorithms

下载图片查看原文

图 5. 基于特征值分解的LSCI滤波算法模型^［64］（ $X$ ′：原始信号； $X$ ：去除白噪声后的图像； $X_{S}$ ：静态散射光信号； $X_{B}$ ：血流信号； $X_{W}$ ：白噪声信号）

Fig. 5. LSCI filtering model based on eigenvalue-decomposition^[64] ( $X^{'}$ : original speckle signal vector; $X$ : speckle signal vector after denoising; $X_{S}$ : static scattered light signal; $X_{B}$ : fluctuating blood signal; $X_{W}$ : white noise signal)

下载图片查看原文

图 6. 基于特征值分解和空间滤波相结合的LSCI滤波算法^［100］

Fig. 6. LSCI filtering algorithm based on eigenvalue-decomposition and filtering^[100]

下载图片查看原文

图 7. 对比实验结果^［100］。（a）原始的眼底衬比图；（b）使用特征值分解和空间滤波处理后的眼底衬比图

Fig. 7. Comparative experimental results^[100]. (a) Raw fundus contrast image; (b) fundus contrast image after eigenvalue-decomposition and spatial filtering

下载图片查看原文

图 8. MD-ABM3D算法模型^［47］

Fig. 8. Scheme of MD-ABM3D algorithm^[47]

下载图片查看原文

图 9. 不同去噪算法的实验结果^［47］。（a）原始图，PSNR为18.5，MSSIM为0.46，R=0.813；（b）savg-tLSCI算法，PSNR为32.8，MSSIM为0.87，R=0.987；（c）NLM算法，PSNR为31.0，MSSIM为0.90，R=0.986；（d）BM3D算法，PSNR为35.8，MSSIM为0.92，R=0.993；（e）MD-ABM3D算法，PSNR为37.8，MSSIM为0.96，R=0.996；（f）参考图

Fig. 9. Output of different denoising algorithms^[47]. (a) Original image, where PSNR is 18.5, MSSIM is 0.46, and R is 0.813; (b) savg-tLSCI algorithm, where PSNR is 32.8, MSSIM is 0.87, and R is 0.987; (c) NLM algorithm, PSNR is 31.0, MSSIM is 0.90, and R is 0.986; (d) BM3D algorithm, PSNR is 35.8, MSSIM is 0.92, and R is 0.993; (e) MD-ABM3D algorithm, PSNR is 37.8, MSSIM is 0.96, and R is 0.996; (f) reference image

下载图片查看原文

图 10. rLASCA算法模型^［61］

Fig. 10. Model of rLASCA algorithm^[61]

下载图片查看原文

图 11. rLASCA算法的实验结果^［61］。（a）未配准的散斑衬比图像；（b）rLASCA算法配准后的散斑衬比图像；（c）图（a）中白色矩形框区域的放大图；（d）图（b）中白色矩形框区域的放大图；（e）白色矩形框区域的白光图

Fig. 11. Experimental results of rLASCA algorithm^[61]. (a) Unregistered laser speckle contrast image; (b) laser speckle contrast image registered by rLASCA; (c) enlarged image of white rectangular box area in figure (a); (d) enlarged image of white rectangular box area in figure (b); (e) white light map of white rectangular box area

下载图片查看原文

图 12. 基于非相干光的非刚体配准算法^［45］。（a）双模态照明装置；（b）算法模型

Fig. 12. Non-rigid registration algorithm based on non-coherent light^[45]. (a) Experimental setup of dual-mode lighting system; (b) algorithm model

下载图片查看原文

图 13. 刚性配准与非刚性配准结果的对比^［45］。（a）未配准的血流图像；（b）血流图像的刚性配准结果；（c）血流图像的非刚性配准结果

Fig. 13. Comparison of rigid registration and non-rigid registration^[45]. (a) Unregistered blood flow image; (b) blood flow image after rigid registration; (c) blood flow image after non-rigid registration

下载图片查看原文

图 14. 基于图像分解的LSCI运动伪影的校正模型^［106］。（a）校正模型；（b）选取回归参量；（c）回归拟合分析；（d）~（f）校正前后的衬比值

Fig. 14. Correction model for LSCI movement artifact based on image decomposition^[106]. (a) Correction model; (b) selection of regression variance; (c) fitted by regression analysis; (d)-(f) contrast value before and after movement correction

下载图片查看原文

图 15. 基于轮廓波变换和多聚焦图像融合算法的LSCI校正模型^［46］

Fig. 15. LSCI correction model based on contourlet transform and multi-focus image fusion^[46]

下载图片查看原文

图 16. 不均匀光强校正前后的结果^［103］。（a）受不均匀性影响的衬比图像；（b）重建后的衬比图像

Fig. 16. Experiment results before and after nonuniform intensity correction^[103]. (a) Contrast image affected by nonuniformity; (b) reconstructed contrast image

下载图片查看原文

图 17. 非均匀校正实验结果^［110］。（a）两种不同光照强度下的灰度散斑图像；（b）从上往下依次是高强度光照、低强度光照下的衬比图以及校正后的低强度光照衬比图；（c）低强度光照、高强度光照以及校正后低强度光照衬比图沿图（a）中横向红线的衬比值曲线；（d）校正后低强度光照衬比图沿图（a）中纵向黄线的衬比值曲线

Fig. 17. Experimental results of nonuniform correction^[110]. (a) Grayscale speckle images at two different intensities; (b) from the top to the bottom: contrast maps at high intensity and low intensity and corrected contrast map at low intensity; (c) contrast profile along the red line marked in figure (a) of contrast maps at low intensity and high intensity and corrected contrast map at low intensity; (d) contrast profile along yellow line marked in figure (a) of corrected contrast map at low intensity

下载图片查看原文

图 18. dLSI算法处理后的血流图像^［84］

Fig. 18. Blood flow image processed by dLSI algorithm^[84]

下载图片查看原文

图 19. 基于多焦点成像的装置示意图^［119］

Fig. 19. Schematic of multi-focus imaging setup^[119]

下载图片查看原文

图 20. DLSI动态散斑衬比校正模型^［74］

Fig. 20. Model of dynamic scattering contrast correction model^[74]

下载图片查看原文

图 21. 空间频域成像LSCI^［121］。（a）si-SFDI实验装置；（b）si-SFDI算法流程

Fig. 21. Spatial frequency domain imaging^{LSCI^[121]. (a) Experimental setup of si-SFDI; (b) processing flow of si-SFDI}

下载图片查看原文

图 22. OSIV成像系统^［10］

Fig. 22. Experimental setup for optical speckle image velocimetry (OSIV)^[10]

下载图片查看原文

图 23. OSIV算法处理流程^［10］

Fig. 23. Processing flow of OSIV algorithm^[10]

下载图片查看原文

图 24. 基于样本熵的衬比分析方法及部分实验结果^［111］。（a）基于样本熵的衬比分析方法；（b）部分实验结果

Fig. 24. Sample entropy-based laser speckle contrast analysis method and partial experimental results^[111]. (a) Sample entropy-based laser speckle contrast analysis method; (b) partial experimental results

下载图片查看原文

图 25. 多曝光激光散斑成像^［83］。（a）MESI系统；（b）单曝光成像和MESI下τ_c的百分比偏差

Fig. 25. Multi-exposure laser speckle imaging^[83]. (a) Multi-exposure speckle imaging system; (b) percentage deviation in $τ_{c}$ under single exposure model and MESI

下载图片查看原文

图 26. 非宽场照明的横向激光散斑对比分析方法^［127］。（a）线性扫描照明的横向激光散斑成像系统；（b）图像处理流程；（c）~（e）传统衬比分析方法、使用常数加权的横向散斑衬比分析方法、使用深度灵敏度曲线加权的横向散斑衬比分析方法获得的血流图像

Fig. 26. Lateral speckle contrast analysis method combined with non-wide field illumination^[127]. (a) Schematic of LSCI experimental setup based on line beam scanning illumination; (b) image processing flow; (c)-(d) blood flow images obtained by traditional contrast analysis method, lateral speckle contrast analysis methods weighted with constant and depth sensitivity curves, respectively

下载图片查看原文

图 27. DSCA成像系统示意图^［132］

Fig. 27. Schematic of DSCA imaging system^[132]

下载图片查看原文

图 28. 激光散斑血流成像系统^［130］。（a）TR-LSCI成像系统；（b）传统的反射式成像系统

Fig. 28. LSCI system for blood flow^[130]. (a) TR-LSCI system; (b) conventional reflective-detected LSCI system

下载图片查看原文

图 29. 新型LSCI系统及其应用研究进展

Fig. 29. Novel LSCI systems and their advances in application and research

下载图片查看原文

图 30. 基于DSP的便携式LSCI系统^［135］。（a）便携式LSCI系统示意图；（b）硬件框架图；（c）软件架构图

Fig. 30. Portable LSCI based on DSP^[135]. (a) Schematic illustration of portable LSCI system; (b) block diagram of hardware framework; (c) block diagram of software framework

下载图片查看原文

图 31. 基于FPGA开发的便携式LSCI系统^［136］

Fig. 31. Portable LSCI based on FPGA^[136]

下载图片查看原文

图 32. 基于嵌入式GPU系统的高性能便携式LSCI系统^［57］

Fig. 32. Efficient portable LSCI based on embedded GPU^[57]

下载图片查看原文

图 33. 内窥式LSCI系统^{［50，88］}

Fig. 33. Endoscopic LSCI system^[50,88]

下载图片查看原文

图 34. 腹腔镜LSCI双显示成像系统^［14］。（a）腹腔镜LSCI成像系统；（b）插入腹腔镜；（c）手持式操作过程；（d）LSCI肠成像；（e）LSCI胆囊成像；（f）LSCI肠系膜成像

Fig. 34. Dual-display laparoscopic laser speckle contrast imaging (LSCI) system^[14]. (a) Laparoscopic LSCI system; (b) inserted laparoscopy; (c) handheld operation; (d) LSCI bowel imaging; (e) LSCI gallbladder imaging; (e) LSCI mesentery imaging

下载图片查看原文

图 35. 头戴式LSCI系统^［60］

Fig. 35. Head-mounted LSCI^[60]

下载图片查看原文

图 36. ECoG-LSCI系统示意图^［23］

Fig. 36. Schematic of ECoG-LSCI^[23]

下载图片查看原文

图 37. 接受电刺激前后大鼠右脑前肢体感皮层卒中区域的激光散斑血流随时间的变化情况^［23］

Fig. 37. Speckle contrast images for rCBF upon electrical stimulation in forelimb- and hindlimb-stimulated groups at serial time points^[23]

下载图片查看原文

图 38. 多模态眼功能成像技术^［17］

Fig. 38. Multimodal and functional imaging of retina^[17]

下载图片查看原文

图 39. 多模态手术实时导航系统^［141］

Fig. 39. Multimodal system for real-time surgical guidance^[141]

下载图片查看原文

表 1动态散斑衬比值校正模型^［115］

Table1. Correction model of dynamic speckle contrast^[115]

Scattering

regime

Velocity

distribution

Speckle visibility expression

(x = T / τ_{c})

Single

Lorentzian

K (T, τ_{c}) = {\{β ρ^{2} \frac{e x p (- 2 x) - 1 + 2 x}{2 x^{2}} + 4 β ρ \frac{e x p (- x) - 1 + x}{x^{2}} + β {(1 - ρ)}^{2} + v_{n o i s e}\}}^{0.5}

Multiple

Gaussian

K (T, τ_{c}) = {\{β ρ^{2} \frac{e x p (- 2 N_{d} x) - 1 + 2 N_{d} x}{2 N_{d}^{2} x^{2}} + 4 β ρ \frac{e x p (- N_{d} x) - 1 + N_{d} x}{N_{d}^{2} x^{2}} + β {(1 - ρ)}^{2} + v_{n o i s e}\}}^{0.5}

查看原文

表 2不同散射特性和粒子运动模式的电场自相关函数^［73］

Table2. Electric field autocorrelation function $g_{1} (τ)$ for different scattering characteristics and particle motion models^[73]

$g_{1} (τ) f o r m$	Scattering regime	Motion	Vessel size	Notation
$e x p (- \sqrt[]{τ / τ_{c}})$	Multiple	Unordered	Small（diameter is about less than $30 μ m$	n=0.5 for MU
$e x p (- τ / τ_{c})$	Multiple	Ordered	Medium （diameter is about 30-110 $μ$ m）	n=1 for MO or SU
$e x p (- τ / τ_{c})$	Single	Unordered	Medium （diameter is about 30-110 $μ$ m）	n=1 for MO or SU
$e x p [- {(τ / τ_{c})}^{2}]$	Single	Ordered	Large（diameter is about more than $110 μ m$ ）	n=2 for SO

查看原文

翟林君, 傅玉青, 杜永兆. 激光散斑衬比血流成像关键技术及应用研究进展[J]. 中国激光, 2023, 50(9): 0907106. Linjun Zhai, Yuqing Fu, Yongzhao Du. Advances in Laser Speckle Contrast Imaging: Key Techniques and Applications[J]. Chinese Journal of Lasers, 2023, 50(9): 0907106.

激光散斑衬比血流成像关键技术及应用研究进展 下载： 1232次

图 1. LSCI实验系统示意图［30］

Fig. 1. Schematic setup for laser speckle contrast imaging (LSCI)[30]

图 2. LSCI关键技术问题的分析及解决思路

Fig. 2. Analysis and solution of key technical problems of LSCI

图 3. aLSCI算法模型［98］

Fig. 3. Scheme of aLSCI algorithm[98]

Fig. 4. Comparative experimental results of different algorithms[98]. (a) tLSCI algorithm; (b) sLSCI algorithm; (c) stLSCI algorithm; (d) savgtLSCI algorithm; (e) tavgsLSCI algorithm; (f) aLSCI algorithm; (g) contrast-to-noise ratio (CNR) of different algorithms

图 5. 基于特征值分解的LSCI滤波算法模型［64］（X′：原始信号；X：去除白噪声后的图像；XS：静态散射光信号；XB：血流信号；XW：白噪声信号）

Fig. 5. LSCI filtering model based on eigenvalue-decomposition[64] (X': original speckle signal vector; X: speckle signal vector after denoising; XS: static scattered light signal; XB: fluctuating blood signal; XW: white noise signal)

图 6. 基于特征值分解和空间滤波相结合的LSCI滤波算法［100］

Fig. 6. LSCI filtering algorithm based on eigenvalue-decomposition and filtering[100]

图 7. 对比实验结果［100］。（a）原始的眼底衬比图；（b）使用特征值分解和空间滤波处理后的眼底衬比图

Fig. 7. Comparative experimental results[100]. (a) Raw fundus contrast image; (b) fundus contrast image after eigenvalue-decomposition and spatial filtering

图 8. MD-ABM3D算法模型［47］

Fig. 8. Scheme of MD-ABM3D algorithm[47]

图 10. rLASCA算法模型［61］

Fig. 10. Model of rLASCA algorithm[61]

图 11. rLASCA算法的实验结果［61］。（a）未配准的散斑衬比图像；（b）rLASCA算法配准后的散斑衬比图像；（c）图（a）中白色矩形框区域的放大图；（d）图（b）中白色矩形框区域的放大图；（e）白色矩形框区域的白光图

图 12. 基于非相干光的非刚体配准算法［45］。（a）双模态照明装置；（b）算法模型

Fig. 12. Non-rigid registration algorithm based on non-coherent light[45]. (a) Experimental setup of dual-mode lighting system; (b) algorithm model

图 13. 刚性配准与非刚性配准结果的对比［45］。（a）未配准的血流图像；（b）血流图像的刚性配准结果；（c）血流图像的非刚性配准结果

Fig. 13. Comparison of rigid registration and non-rigid registration[45]. (a) Unregistered blood flow image; (b) blood flow image after rigid registration; (c) blood flow image after non-rigid registration

图 14. 基于图像分解的LSCI运动伪影的校正模型［106］。（a）校正模型；（b）选取回归参量；（c）回归拟合分析；（d）~（f）校正前后的衬比值

Fig. 14. Correction model for LSCI movement artifact based on image decomposition[106]. (a) Correction model; (b) selection of regression variance; (c) fitted by regression analysis; (d)-(f) contrast value before and after movement correction

图 15. 基于轮廓波变换和多聚焦图像融合算法的LSCI校正模型［46］

Fig. 15. LSCI correction model based on contourlet transform and multi-focus image fusion[46]

图 16. 不均匀光强校正前后的结果［103］。（a）受不均匀性影响的衬比图像；（b）重建后的衬比图像

Fig. 16. Experiment results before and after nonuniform intensity correction[103]. (a) Contrast image affected by nonuniformity; (b) reconstructed contrast image

图 18. dLSI算法处理后的血流图像［84］

Fig. 18. Blood flow image processed by dLSI algorithm[84]

图 19. 基于多焦点成像的装置示意图［119］

Fig. 19. Schematic of multi-focus imaging setup[119]

图 20. DLSI动态散斑衬比校正模型［74］

Fig. 20. Model of dynamic scattering contrast correction model[74]

图 21. 空间频域成像LSCI［121］。（a）si-SFDI实验装置；（b）si-SFDI算法流程

Fig. 21. Spatial frequency domain imagingLSCI[121]. (a) Experimental setup of si-SFDI; (b) processing flow of si-SFDI

图 22. OSIV成像系统［10］

Fig. 22. Experimental setup for optical speckle image velocimetry (OSIV)[10]

图 23. OSIV算法处理流程［10］

Fig. 23. Processing flow of OSIV algorithm[10]

图 24. 基于样本熵的衬比分析方法及部分实验结果［111］。（a）基于样本熵的衬比分析方法；（b）部分实验结果

Fig. 24. Sample entropy-based laser speckle contrast analysis method and partial experimental results[111]. (a) Sample entropy-based laser speckle contrast analysis method; (b) partial experimental results

图 25. 多曝光激光散斑成像［83］。（a）MESI系统；（b）单曝光成像和MESI下τc的百分比偏差

Fig. 25. Multi-exposure laser speckle imaging[83]. (a) Multi-exposure speckle imaging system; (b) percentage deviation in τc under single exposure model and MESI

图 27. DSCA成像系统示意图［132］

Fig. 27. Schematic of DSCA imaging system[132]

图 28. 激光散斑血流成像系统［130］。（a）TR-LSCI成像系统；（b）传统的反射式成像系统

Fig. 28. LSCI system for blood flow[130]. (a) TR-LSCI system; (b) conventional reflective-detected LSCI system

图 29. 新型LSCI系统及其应用研究进展

Fig. 29. Novel LSCI systems and their advances in application and research

图 30. 基于DSP的便携式LSCI系统［135］。（a）便携式LSCI系统示意图；（b）硬件框架图；（c）软件架构图

Fig. 30. Portable LSCI based on DSP[135]. (a) Schematic illustration of portable LSCI system; (b) block diagram of hardware framework; (c) block diagram of software framework

图 31. 基于FPGA开发的便携式LSCI系统［136］

Fig. 31. Portable LSCI based on FPGA[136]

图 32. 基于嵌入式GPU系统的高性能便携式LSCI系统［57］

Fig. 32. Efficient portable LSCI based on embedded GPU[57]

图 33. 内窥式LSCI系统［50，88］

Fig. 33. Endoscopic LSCI system[50,88]

图 34. 腹腔镜LSCI双显示成像系统［14］。（a）腹腔镜LSCI成像系统；（b）插入腹腔镜；（c）手持式操作过程；（d）LSCI肠成像；（e）LSCI胆囊成像；（f）LSCI肠系膜成像

Fig. 34. Dual-display laparoscopic laser speckle contrast imaging (LSCI) system[14]. (a) Laparoscopic LSCI system; (b) inserted laparoscopy; (c) handheld operation; (d) LSCI bowel imaging; (e) LSCI gallbladder imaging; (e) LSCI mesentery imaging

图 35. 头戴式LSCI系统［60］

Fig. 35. Head-mounted LSCI[60]

图 36. ECoG-LSCI系统示意图［23］

Fig. 36. Schematic of ECoG-LSCI[23]

图 37. 接受电刺激前后大鼠右脑前肢体感皮层卒中区域的激光散斑血流随时间的变化情况［23］

Fig. 37. Speckle contrast images for rCBF upon electrical stimulation in forelimb- and hindlimb-stimulated groups at serial time points[23]

图 38. 多模态眼功能成像技术［17］

Fig. 38. Multimodal and functional imaging of retina[17]

图 39. 多模态手术实时导航系统［141］

Fig. 39. Multimodal system for real-time surgical guidance[141]

表 1动态散斑衬比值校正模型［115］

Table1. Correction model of dynamic speckle contrast[115]

表 2不同散射特性和粒子运动模式的电场自相关函数［73］

Table2. Electric field autocorrelation function g1τ for different scattering characteristics and particle motion models[73]

相关论文

相关资讯

关于本站 Cookie 的使用提示

全站搜索

激光散斑衬比血流成像关键技术及应用研究进展下载： 1232次

图 1. LSCI实验系统示意图^［30］

Fig. 1. Schematic setup for laser speckle contrast imaging (LSCI)^[30]

图 3. aLSCI算法模型^［98］

Fig. 3. Scheme of aLSCI algorithm^[98]

Fig. 4. Comparative experimental results of different algorithms^[98]. (a) tLSCI algorithm; (b) sLSCI algorithm; (c) stLSCI algorithm; (d) savgtLSCI algorithm; (e) tavgsLSCI algorithm; (f) aLSCI algorithm; (g) contrast-to-noise ratio (CNR) of different algorithms

图 5. 基于特征值分解的LSCI滤波算法模型^［64］（ $X$ ′：原始信号； $X$ ：去除白噪声后的图像； $X_{S}$ ：静态散射光信号； $X_{B}$ ：血流信号； $X_{W}$ ：白噪声信号）

Fig. 5. LSCI filtering model based on eigenvalue-decomposition^[64] ( $X^{'}$ : original speckle signal vector; $X$ : speckle signal vector after denoising; $X_{S}$ : static scattered light signal; $X_{B}$ : fluctuating blood signal; $X_{W}$ : white noise signal)

图 6. 基于特征值分解和空间滤波相结合的LSCI滤波算法^［100］

Fig. 6. LSCI filtering algorithm based on eigenvalue-decomposition and filtering^[100]

图 7. 对比实验结果^［100］。（a）原始的眼底衬比图；（b）使用特征值分解和空间滤波处理后的眼底衬比图

Fig. 7. Comparative experimental results^[100]. (a) Raw fundus contrast image; (b) fundus contrast image after eigenvalue-decomposition and spatial filtering

图 8. MD-ABM3D算法模型^［47］

Fig. 8. Scheme of MD-ABM3D algorithm^[47]

图 10. rLASCA算法模型^［61］

Fig. 10. Model of rLASCA algorithm^[61]

图 11. rLASCA算法的实验结果^［61］。（a）未配准的散斑衬比图像；（b）rLASCA算法配准后的散斑衬比图像；（c）图（a）中白色矩形框区域的放大图；（d）图（b）中白色矩形框区域的放大图；（e）白色矩形框区域的白光图

图 12. 基于非相干光的非刚体配准算法^［45］。（a）双模态照明装置；（b）算法模型

Fig. 12. Non-rigid registration algorithm based on non-coherent light^[45]. (a) Experimental setup of dual-mode lighting system; (b) algorithm model

图 13. 刚性配准与非刚性配准结果的对比^［45］。（a）未配准的血流图像；（b）血流图像的刚性配准结果；（c）血流图像的非刚性配准结果

Fig. 13. Comparison of rigid registration and non-rigid registration^[45]. (a) Unregistered blood flow image; (b) blood flow image after rigid registration; (c) blood flow image after non-rigid registration

图 14. 基于图像分解的LSCI运动伪影的校正模型^［106］。（a）校正模型；（b）选取回归参量；（c）回归拟合分析；（d）~（f）校正前后的衬比值

Fig. 14. Correction model for LSCI movement artifact based on image decomposition^[106]. (a) Correction model; (b) selection of regression variance; (c) fitted by regression analysis; (d)-(f) contrast value before and after movement correction

图 15. 基于轮廓波变换和多聚焦图像融合算法的LSCI校正模型^［46］

Fig. 15. LSCI correction model based on contourlet transform and multi-focus image fusion^[46]

图 16. 不均匀光强校正前后的结果^［103］。（a）受不均匀性影响的衬比图像；（b）重建后的衬比图像

Fig. 16. Experiment results before and after nonuniform intensity correction^[103]. (a) Contrast image affected by nonuniformity; (b) reconstructed contrast image

图 18. dLSI算法处理后的血流图像^［84］

Fig. 18. Blood flow image processed by dLSI algorithm^[84]

图 19. 基于多焦点成像的装置示意图^［119］

Fig. 19. Schematic of multi-focus imaging setup^[119]

图 20. DLSI动态散斑衬比校正模型^［74］

Fig. 20. Model of dynamic scattering contrast correction model^[74]

图 21. 空间频域成像LSCI^［121］。（a）si-SFDI实验装置；（b）si-SFDI算法流程

Fig. 21. Spatial frequency domain imaging^{LSCI^[121]. (a) Experimental setup of si-SFDI; (b) processing flow of si-SFDI}

图 22. OSIV成像系统^［10］

Fig. 22. Experimental setup for optical speckle image velocimetry (OSIV)^[10]

图 23. OSIV算法处理流程^［10］

Fig. 23. Processing flow of OSIV algorithm^[10]

图 24. 基于样本熵的衬比分析方法及部分实验结果^［111］。（a）基于样本熵的衬比分析方法；（b）部分实验结果

Fig. 24. Sample entropy-based laser speckle contrast analysis method and partial experimental results^[111]. (a) Sample entropy-based laser speckle contrast analysis method; (b) partial experimental results

图 25. 多曝光激光散斑成像^［83］。（a）MESI系统；（b）单曝光成像和MESI下τ_c的百分比偏差

Fig. 25. Multi-exposure laser speckle imaging^[83]. (a) Multi-exposure speckle imaging system; (b) percentage deviation in $τ_{c}$ under single exposure model and MESI

图 27. DSCA成像系统示意图^［132］

Fig. 27. Schematic of DSCA imaging system^[132]

图 28. 激光散斑血流成像系统^［130］。（a）TR-LSCI成像系统；（b）传统的反射式成像系统

Fig. 28. LSCI system for blood flow^[130]. (a) TR-LSCI system; (b) conventional reflective-detected LSCI system

图 30. 基于DSP的便携式LSCI系统^［135］。（a）便携式LSCI系统示意图；（b）硬件框架图；（c）软件架构图

Fig. 30. Portable LSCI based on DSP^[135]. (a) Schematic illustration of portable LSCI system; (b) block diagram of hardware framework; (c) block diagram of software framework

图 31. 基于FPGA开发的便携式LSCI系统^［136］

Fig. 31. Portable LSCI based on FPGA^[136]

图 32. 基于嵌入式GPU系统的高性能便携式LSCI系统^［57］

Fig. 32. Efficient portable LSCI based on embedded GPU^[57]

图 33. 内窥式LSCI系统^{［50，88］}

Fig. 33. Endoscopic LSCI system^[50,88]

图 34. 腹腔镜LSCI双显示成像系统^［14］。（a）腹腔镜LSCI成像系统；（b）插入腹腔镜；（c）手持式操作过程；（d）LSCI肠成像；（e）LSCI胆囊成像；（f）LSCI肠系膜成像

Fig. 34. Dual-display laparoscopic laser speckle contrast imaging (LSCI) system^[14]. (a) Laparoscopic LSCI system; (b) inserted laparoscopy; (c) handheld operation; (d) LSCI bowel imaging; (e) LSCI gallbladder imaging; (e) LSCI mesentery imaging

图 35. 头戴式LSCI系统^［60］

Fig. 35. Head-mounted LSCI^[60]

图 36. ECoG-LSCI系统示意图^［23］

Fig. 36. Schematic of ECoG-LSCI^[23]

图 37. 接受电刺激前后大鼠右脑前肢体感皮层卒中区域的激光散斑血流随时间的变化情况^［23］

Fig. 37. Speckle contrast images for rCBF upon electrical stimulation in forelimb- and hindlimb-stimulated groups at serial time points^[23]

图 38. 多模态眼功能成像技术^［17］

Fig. 38. Multimodal and functional imaging of retina^[17]

图 39. 多模态手术实时导航系统^［141］

Fig. 39. Multimodal system for real-time surgical guidance^[141]

表 1动态散斑衬比值校正模型^［115］

Table1. Correction model of dynamic speckle contrast^[115]

表 2不同散射特性和粒子运动模式的电场自相关函数^［73］

Table2. Electric field autocorrelation function $g_{1} (τ)$ for different scattering characteristics and particle motion models^[73]