CC BY
6*
ALT'23
The 30th International Conference on Advanced Laser Technologies
B-I-14
Motion correction of laser speckle imaging of blood flow
Xiaohu Liu1, Liangwei Meng1, Jinling Lu1, Pengcheng Li1'2*
1-Britton Chance Center for Biomedical Photonics and MoE Key Laboratory for Biomedical Photonics, Wuhan National Laboratory for Optoelectronics, Huazhong University of Science and Technology, China 2- Department of Biomedical Engineering, Hainan University, China *Corresponding author: [email protected]
Laser speckle contrast imaging (LSCI) is widely used for mapping the blood flow speed of biological tissue. Since multiple speckle images are usually required to reconstruct a high-resolution and high SNR blood flow map, LSCI is highly sensitive to the motion artifacts. The blood flow perfusion obtained by LSCI during tissue motion is a mixture of the artifacts due to inter-frame image misalignment and intra-frame image blurring. It is of urgent importance to minimize the impact of motion when LSCI is put into practical use. Previous studies have mainly focused on correcting either the inter-frame or intra-frame artifact of global rigid motion. However, in practical biomedical applications, owing to the flexibility of biological tissues, the subjects often produce many non- rigid, spatially non-uniform movements, such as respiration, heartbeats, peristalsis of digestive organs. Such artifacts cannot be well corrected only with an estimation of global movement of tissues and rigid registration. In order to solve this problem, it is essential to accurately detect the heterogeneous local movements of biological tissues in the field of view and perform non-rigid registration. It is difficult to correct the local non-rigid motion only by using granular raw speckle images, especially in the tissue dominated by small blood vessels. We proposed a dual-wavelength imaging system to solve the problem that laser speckle images with graininess are difficult to be directly registered. Laser speckle images and corresponding tissue structure images are acquired simultaneously by a color CMOS camera under the illumination of 660 nm laser diode and a 470 nm light-emitting diode. The registration matrix is extracted from the tissue structure images registration process, which is then used to register the corresponding laser speckle images [1]. Considering the graininess of the speckle images, the application of the nearest neighbor interpolation method is proposed during registration to preserve the original speckle information, so as to avoid the overestimation of the blood flow speed. To further address the problem of overestimation of blood flow due to non-rigid tissue motion during the exposure of a single frame, we proposed a method that combines the strategy of heterogeneous regression analysis with the above dual-wavelength imaging, enabling the extraction of the true blood flow signal from the measured mixture LSCI signal.
[1] X. Liu, J. Wei, L. Meng, W. Cheng, X. Zhu, J. Lu and P. Li. Motion correction of laser speckle contrast imaging of blood flow by simultaneous
imaging of tissue structure and non-rigid registration, Optics and Lasers in Engineering, Vol. 140, pp.106526, (2021)
