Multispectral Imaging for Vein Localization and Contrast Enhancement.

Intravenous (IV) catheterization is a common procedure. Still, there is a 26% chance of the first attempt catheterization failure due to the changing visibility of veins because of the patient's skin tone and body fat content. Ultrasound assistive devices help locate deeper veins but are not practical in emergencies, and transillumination assistive devices have a low field of view. Commercial near-infrared (NIR) imaging devices are effective in vein localization but are expensive and are not used in low-cost clinical settings. To overcome this, NIR Multispectral Imaging (MSI) was used to find the optimal wavelength that provides the enhanced visualization of veins for all skin types and Body Mass Index (BMI). The band with the highest vein-to-skin contrast ratio was selected and contrast enhancement was done using our proposed method. The primary blocks of the proposed method are Gamma correction, Contrast Limited Adaptive Histogram Equalization (CLAHE), Adaptive Thresholding, and image Fusion. The optimal spectral range was found to be 814-876 nm and our method increased the contrast by 0.41, 0.375, and 0.39 for fair, brown, and dark brown skin types, respectively, with different BMI.Clinical relevance- From the study, we can develop a potentially low-cost vein localization assistive device for training medical and nursing students and use it in emergencies for venous access to improve confidence in IV catheterization.

A Multifunctional Network with Uncertainty Estimation and Attention-Based Knowledge Distillation to Address Practical Challenges in Respiration Rate Estimation.

Respiration rate is a vital parameter to indicate good health, wellbeing, and performance. As the estimation through classical measurement modes are limited only to rest or during slow movements, respiration rate is commonly estimated through physiological signals such as electrocardiogram and photoplethysmography due to the unobtrusive nature of wearable devices. Deep learning methodologies have gained much traction in the recent past to enhance accuracy during activities involving a lot of movement. However, these methods pose challenges, including model interpretability, uncertainty estimation in the context of respiration rate estimation, and model compactness in terms of deployment in wearable platforms. In this direction, we propose a multifunctional framework, which includes the combination of an attention mechanism, an uncertainty estimation functionality, and a knowledge distillation framework. We evaluated the performance of our framework on two datasets containing ambulatory movement. The attention mechanism visually and quantitatively improved instantaneous respiration rate estimation. Using Monte Carlo dropouts to embed the network with inferential uncertainty estimation resulted in the rejection of 3.7% of windows with high uncertainty, which consequently resulted in an overall reduction of 7.99% in the mean absolute error. The attention-aware knowledge distillation mechanism reduced the model's parameter count and inference time by 49.5% and 38.09%, respectively, without any increase in error rates. Through experimentation, ablation, and visualization, we demonstrated the efficacy of the proposed framework in addressing practical challenges, thus taking a step towards deployment in wearable edge devices.

Comparative Analysis of Resting Heart Rate Measurement at Multiple Instances in a Single Day.

Resting Heart Rate (RHR) is used as an indicator of cardiovascular health and overall fitness. Clinically, RHR is measured from beat-to-beat heart rate data during the day when the body is at rest (RHRrest), typically for ≥ 5 minutes. In this paper, we have compared the RHR measurements done at multiple instances in a single day namely, [Formula: see text], RHR immediately after waking up (RHRmorning) and RHR during sleep (RHRsleep). The significance of measuring RHRsleep and why it can be used as a potential replacement for the conventional methods is analysed through an experimental study in this paper. The results obtained using the proposed method stands out in terms of repeatability. RHR measurements were taken for 3 instances on a single day for 9 subjects on 5 alternate workdays. A comparative analysis was performed by measuring the repeatability coefficient (RC) and Standard Deviation (SD) on the RHR measurements taken during multiple instances for each subject separately. The average RC and SD over the 5 alternate workdays was 5 bpm and SD was 2 bpm for RHRslep. For RHRrest and RHRmorning, the average RC was 12 bpm and 11 bpm and the average SD was 5 bpm and 4 bpm respectively, which is comparatively higher. Hence this method can be potentially adopted instead of the conventional methods as the RHRsleep parameter is more reliable and precise due to its repeatable nature.

Enhanced Vascular Features in Porcine Gastrointestinal Endoscopy Using Multispectral Imaging.

Endoscopic investigation is a predominant stan-dard while assessing the gastrointestinal tract. Even though it has been rigorously used in diagnostics for many decades, a high miss rate has been recorded. Advanced endoscopic imaging still has not found solutions to problems like early cancer detection, polyp generality, disease classification, etc. One of the less explored techniques to study early cancer detection is spectral imaging which deals with the absorption and reflection spectra of various wavelengths of light by different layers of tissue. To study tissues under various illumination, a multi-spectral light source unit that can be used along with an endoscopy system was developed with 10 different LEDs of very narrow bandwidths. Using this light source, a feasibility study was per-formed on an animal in which the upper GI tract of a porcine model was imaged and sample images were taken for processing from five different sections. Some wavelengths showed better contrast enhancements for visualization of vascular structures. Wavelength 420 nm (violet light) showed better contrast and the gradient of the line profile histogram showed the highest intensity change between the blood vessels and the surrounding mucosa. These enhancements showed that spectral imaging can potentially help in studying tissues for early cancer detection and improved visualization of the G I tract using endoscopy.

RPnet: A Deep Learning approach for robust R Peak detection in noisy ECG.

Automatic detection of R-peaks in an Electrocardiogram signal is crucial in a multitude of applications including Heart Rate Variability (HRV) analysis and Cardio Vascular Disease(CVD) diagnosis. Although there have been numerous approaches that have successfully addressed the problem, there has been a notable dip in the performance of these existing detectors on ECG episodes that contain noise and HRV Irregulates. On the other hand, Deep Learning(DL) based methods have shown to be adept at modelling data that contain noise. In image to image translation, Unet is the fundamental block in many of the networks. In this work, a novel application of the Unet combined with Inception and Residual blocks is proposed to perform the extraction of R-peaks from an ECG. Furthermore, the problem formulation also robustly deals with issues of variability and sparsity of ECG R-peaks. The proposed network was trained on a database containing ECG episodes that have CVD and was tested against three traditional ECG detectors on a validation set. The model achieved an F1 score of 0.9837, which is a substantial improvement over the other beat detectors. Furthermore, the model was also evaluated on three other databases. The proposed network achieved high F1 scores across all datasets which established its generalizing capacity. Additionally, a thorough analysis of the model's performance in the presence of different levels of noise was carried out.

Robust Modelling of Reflectance Pulse Oximetry for SpO2 Estimation.

Continuous monitoring of blood oxygen saturation levels is vital for patients with pulmonary disorders. Traditionally, SpO2 monitoring has been carried out using transmittance pulse oximeters due to its dependability. However, SpO2 measurement from transmittance pulse oximeters is limited to peripheral regions. This becomes a disadvantage at very low temperatures as blood perfusion to the peripherals decreases. On the other hand, reflectance pulse oximeters can be used at various sites like finger, wrist, chest and forehead. Additionally, reflectance pulse oximeters can be scaled down to affordable patches that do not interfere with the user's diurnal activities. However, accurate SpO2 estimation from reflectance pulse oximeters is challenging due to its patient dependent, subjective nature of measurement. Recently, a Machine Learning (ML) method was used to model reflectance waveforms onto SpO2 obtained from transmittance waveforms. However, the generalizability of the model to new patients was not tested. In light of this, the current work implemented multiple ML based approaches which were subsequently found to be incapable of generalizing to new patients. Furthermore, a minimally calibrated data driven approach was utilized in order to obtain SpO2 from reflectance PPG waveforms. The proposed solution produces an average mean absolute error of 1.81% on unseen patients which is well within the clinically permissible error of 2%. Two statistical tests were conducted to establish the effectiveness of the proposed method.Clinical relevance- The proposed method ameliorates our current understanding of reflectance based pulse oximetry and provides a method to estimate SpO2 from reflectance pulse oximeters.