Speed-resolved perfusion imaging using multi-exposure laser speckle contrast imaging and machine learning

Enhancing blood flow prediction in multi-exposure laser speckle contrast imaging through ensemble learning with K-mean clustering

Purpose.Accurately visualizing and measuring blood flow is of utmost importance in maintaining optimal health and preventing the onset of various chronic diseases. One promising imaging technique that aids in visualizing perfusion in biological tissues is Multi-exposure Laser Speckle Contrast Imaging (MELSCI). MELSCI technique allows real-time quantitative measurements using multiple exposure times to obtain precise and reliable blood flow data. Additionally, the application of machine learning (ML) techniques can further enhance the accuracy of blood flow prediction in this imaging modality.Method.Our study focused on developing and evaluating Ensemble Learning ML techniques along with clustering algorithms for predicting blood flow rates in MELSCI. The effectiveness of these techniques was assessed using performance parameters, including accuracy, F1-score, precision, recall, specificity, and classification error rate.Result.Notably, the study revealed that Ensemble Learning with clustering emerged as the most accurate technique, achieving an impressive accuracy rate of 98.5%. Furthermore, it demonstrated a high recall of more than 91%, F1-score, the precision of more than 90%, higher specificity of 99%, and least classification error of 1.5%, highlighting its suitability and sustainability for flow prediction in MELSCI.Conclusion.The study's findings imply that Ensemble Learning can significantly contribute to enhancing the accuracy of blood flow prediction in MELSCI. This advancement holds substantial promise for healthcare professionals and researchers, as it facilitates improved understanding and assessment of perfusion within biological tissues, which will contribute to the maintenance of good health and prevention of chronic diseases.

Multi-exposure Laser Speckle Contrast Imaging (MECI)-Based Prediction of Blood Flow Using Random Forest (RF) With K-Means (KM)

Blood flow prediction is very important for medical diagnosis, drug development, tissue engineering, and continuous monitoring. One commonly used method for studying blood flow is called multi-exposure laser speckle contrast imaging (MECI). It provides valuable insights into how blood flows through tissues and helps in diagnosing circulatory diseases. In our study, we used MECI to measure blood flow in real-time by taking multiple measurements with different exposure times and contrasts. To predict different blood flow rates ranging from 0.1 to 1 mm/s, we employed machine learning (ML) techniques like clustering and random forest (RF) or support vector machine (SVM) algorithms. The study showed that RF with K-means performance is found to be the most accurate technique for flow classification, with an accuracy of 98.5%, a precision of 92%, a specificity of 98.9%, and a classification error of 1.5%. Our study demonstrates that employing clustering and RF algorithms in MECI provides a robust and effective approach to predicting blood flow. This technique holds great potential for a wide range of applications in the medical and healthcare fields.

Prediction of motion artifacts caused by translation in handheld laser speckle contrast imaging

Significance

In handheld laser speckle contrast imaging (LSCI), motion artifacts (MA) are inevitable. Suppression of MA leads to a valid and objective assessment of tissue perfusion in a wide range of medical applications including dermatology and burns. Our study shines light on the sources of these artifacts, which have not yet been explored. We propose a model based on optical Doppler effect to predict speckle contrast drop as an indication of MA.

Aim

We aim to theoretically model MA when an LSCI system measuring on static scattering media is subject to translational displacements. We validate the model using both simulation and experiments. This is the crucial first step toward creating robustness against MA.

Approach

Our model calculates optical Doppler shifts in order to predict intensity correlation function and contrast of the time-integrated intensity as functions of applied speed based on illumination and detection wavevectors. To validate the theoretical predictions, computer simulation of the dynamic speckles has been carried out. Then experiments are performed by both high-speed and low-framerate imaging. The employed samples for the experiments are a highly scattering matte surface and a Delrin plate of finite scattering level in which volume scattering occurs.

Results

An agreement has been found between theoretical prediction, simulation, and experimental results of both intensity correlation functions and speckle contrast. Coefficients in the proposed model have been linked to the physical parameters according to the experimental setups.

Conclusions

The proposed model provides a quantitative description of the influence of the types of illumination and media in the creation of MA. The accurate prediction of MA caused by translation based on Doppler shifts makes our model suitable to study the influence of rotation. Also the model can be extended for the case of dynamic media, such as live tissue.