Body Landmarks Research Articles

Diagnosis of Pregnancy Complications Using Blind Ultrasound Sweeps Performed by Individuals Without Prior Formal Ultrasound Training.

To estimate the diagnostic accuracy of blind ultrasound sweeps performed with a low-cost, portable ultrasound system by individuals with no prior formal ultrasound training to diagnose common pregnancy complications. This is a single-center, prospective cohort study conducted from October 2020 to January 2022 among people with second- and third-trimester pregnancies. Nonspecialists with no prior formal ultrasound training underwent a brief training on a simple eight-step approach to performing a limited obstetric ultrasound examination that uses blind sweeps of a portable ultrasound probe based on external body landmarks. The sweeps were interpreted by five blinded maternal-fetal medicine subspecialists. Sensitivity, specificity, and positive and negative predictive values for blinded ultrasound sweep identification of pregnancy complications (fetal malpresentation, multiple gestations, placenta previa, and abnormal amniotic fluid volume) were compared with a reference standard ultrasonogram as the primary analysis. Kappa for agreement was also assessed. Trainees performed 194 blinded ultrasound examinations on 168 unique pregnant people (248 fetuses) at a mean of 28±5.85 weeks of gestation for a total of 1,552 blinded sweep cine clips. There were 49 ultrasonograms with normal results (control group) and 145 ultrasonograms with abnormal results with known pregnancy complications. In this cohort, the sensitivity for detecting a prespecified pregnancy complication was 91.7% (95% CI 87.2-96.2%) overall, with the highest detection rate for multiple gestations (100%, 95% CI 100-100%) and noncephalic presentation (91.8%, 95% CI 86.4-97.3%). There was high negative predictive value for placenta previa (96.1%, 95% CI 93.5-98.8%) and abnormal amniotic fluid volume (89.5%, 95% CI 85.3-93.6%). There was also substantial to perfect mean agreement for these same outcomes (range 87-99.6% agreement, Cohen κ range 0.59-0.91, P<.001 for all). Blind ultrasound sweeps of the gravid abdomen guided by an eight-step protocol using only external anatomic landmarks and performed by previously untrained operators with a low-cost, portable, battery-powered device had excellent sensitivity and specificity for high-risk pregnancy complications such as malpresentation, placenta previa, multiple gestations, and abnormal amniotic fluid volume, similar to results of a diagnostic ultrasound examination using a trained ultrasonographer and standard-of-care ultrasound machine. This approach has the potential to improve access to obstetric ultrasonography globally.

Blood flow restricted walking alters gait kinematics

ABSTRACT This study investigated the impact of blood flow restriction (BFR) during treadmill walking on gait kinematics. Twenty-one participants completed one familiarisation and four experimental sessions, including two walking speeds (moderate [5.0 ± 0.3km·h−1] and fast [6.4 ± 0.4km·h−1]) and two occlusion conditions (BFR [60% of arterial occlusion pressure] and unrestricted). For each exercise intensity, the BFR session was performed first. Participants were instructed to walk as long as possible, with sessions capped at 20 min. Unrestricted sessions were time-matched, and the order of exercise intensity was randomised. Kinematics were collected over 10s every minute using retro-reflective markers affixed to specific body landmarks. Ratings of perceived exertion and discomfort were collected every two minutes. Blood samples were collected from the fingertip pre-exercise and the finger and toe post-exercise, and were analysed for lactate, electrolytes, and markers of cell-membrane damage. During the BFR sessions the cuffs remained inflated while the blood samples were collected. Fast-walk BFR sessions exhibited higher anterior trunk flexion (p = 0.001) and knee flexion during stance (p = 0.001) compared to all other sessions. Step width was increased during BFR sessions (p = 0.001), but no difference in step length (p = 0.300) or cadence (p = 0.922) were observed. The time required to elicit change in anterior trunk flexion and plantar-flexion angle at toe-off was shorter during BFR sessions (p = 0.024). The BFR sessions elicited the highest ratings of perceived exertion and discomfort, as well as blood lactate concentration (p ≤ 0.001). Application of BFR during moderate and fast treadmill walking modifies gait kinematics and exacerbates exercise-related sensations as well as blood lactate concentration.

Deep learning based landmark detection for measuring hock and knee angles in sows

Abstract This paper presents a visual deep learning approach to automatically determine hock and knee angles from sow images. Lameness is the second largest reason for culling of breeding herd females and relies on human observers to provide visual scoring for detection which can be slow, subjective, and inconsistent. A deep learning model classified and detected ten and two key body landmarks from the side and rear profile images, respectively (mean average precision = 0.94). Trigonometric-based formulae were derived to calculate hock and knee angles using the features extracted from the imagery. Automated angle measurements were compared with manual results from each image (average root mean square error [RMSE] = 4.13°), where all correlation slopes (average R2 = 0.84) were statistically different from zero (P &amp;lt; 0.05); all automated measurements were in statistical agreement with manually collected measurements using the Bland–Altman procedure. This approach will be of interest to animal geneticists, scientists, and practitioners for obtaining objective angle measurements that can be factored into gilt replacement criteria to optimize sow breeding units.

Three-dimensional unsupervised probabilistic pose reconstruction (3D-UPPER) for freely moving animals

A key step in understanding animal behaviour relies in the ability to quantify poses and movements. Methods to track body landmarks in 2D have made great progress over the last few years but accurate 3D reconstruction of freely moving animals still represents a challenge. To address this challenge here we develop the 3D-UPPER algorithm, which is fully automated, requires no a priori knowledge of the properties of the body and can also be applied to 2D data. We find that 3D-UPPER reduces by >10 fold the error in 3D reconstruction of mouse body during freely moving behaviour compared with the traditional triangulation of 2D data. To achieve that, 3D-UPPER performs an unsupervised estimation of a Statistical Shape Model (SSM) and uses this model to constrain the viable 3D coordinates. We show, by using simulated data, that our SSM estimator is robust even in datasets containing up to 50% of poses with outliers and/or missing data. In simulated and real data SSM estimation converges rapidly, capturing behaviourally relevant changes in body shape associated with exploratory behaviours (e.g. with rearing and changes in body orientation). Altogether 3D-UPPER represents a simple tool to minimise errors in 3D reconstruction while capturing meaningful behavioural parameters.

View-Invariant Center-of-Pressure Metrics Estimation With Monocular RGB Camera

Center of pressure (CoP) metrics, including CoP path length and sway area, have been used as gold standard measurements of postural and balance control in biomechanical studies. A recent study of computer-vision-based CoP metrics estimation from 3D body landmark sequences offers a more portable and comprehensive solution than conventional force plate methods to obtain these important metrics for real-time evaluation of balance control. However, obtaining accurate 3D body landmarks requires a calibrated motion capture system or on-body markers, which involves lengthy data collection and processing time and limits their implementation in home and clinical environments. Existing methods that instead use 2D body landmarks fail to adapt to different camera positions. To overcome these challenges, we propose a view-invariant deep learning framework for video-level CoP metrics estimation, including CoP path length and sway area, using pose dimension lifting and graph convolutional network (GCN). This work is the first step toward obtaining gold-standard CoP metrics with an accessible, monocular RGB camera. We propose to use a dimension lifting convolutional neural network (CNN) to obtain view-invariant 3D body landmark features from 2D body landmarks. We also propose a two-stream regression model using GCN and discrete cosine transform (DCT) for a robust CoP metrics estimation. To facilitate the line of research, we release a novel multi-view body landmark dataset containing 2D body landmarks of a wide variety of action patterns from four different camera views with synchronized CoP labels and corresponding 3D body landmarks, which enables cross-view evaluation with different camera angles. We subsequently validate the proposed method through a cross-dataset training by training the dimension lifting model on an existing balance dataset and evaluating the CoP metrics estimation on the multi-view body landmark dataset. The experiments validate that our framework achieves state-of-the-art accuracy for both CoP path length and CoP sway area using a monocular RGB camera input for unseen views.

Equine Kinematic Gait Analysis Using Stereo Videography and Deep Learning: Stride Length and Stance Duration Estimation

Highlights Stereo machine vision and deep learning techniques were investigated for infield equine kinematic gait analysis. The proposed pipeline tracks equine body landmarks in 3D space and estimates stride length and stance duration. The system can serve as a cost-effective, rapid, and easy-to-use tool for equine locomotion research. Abstract. Equine kinematic gait analysis (EKGA) currently requires a complicated, expensive, and labor-intensive procedure for equine locomotion research. An automated stereo video processing pipeline was developed and evaluated for measuring equine biomechanical parameters. Using stereo videos of 40 different walking horses, a DeepLabCut (DLC) model was trained to detect body landmarks in individual frames. With an autoregressive integrated moving average filter, the landmark detection had a root mean square error of 5.14 pixels and a mean absolute error of 4.87 pixels. As a case study, methods were developed to extract stride length (SL) and stance duration (SD). Individual hoof gait phase detection was achieved using a fine-tuned Faster R-CNN model and a mode filter, yielding precision and recall values of 0.83 and 0.95, respectively. The semi-global block matching (SGBM) algorithm was used to estimate depth maps, and the accuracy was assessed by comparing head length estimation with infield measurements. A Bland-Altman analysis for DLC-detected head length in combination with SGBM-based 3D reconstruction yielded a bias of -0.014 m with upper and lower limits of agreement (LoAs) of 0.03 m and -0.061 m, respectively. Furthermore, Bland-Altman analyses on SD and SL when compared to image-level manual measurements showed biases of -0.02 sec and -0.042 m, respectively. The corresponding LoAs were (0.01907 sec, -0.24 sec) for SD and (0.04 m, -0.12 m) for SL. The proposed method showed promising potential in performing EKGA in an automated, cost-effective, and rapid manner under field conditions. Keywords: 3D Reconstruction, Animal Pose Estimation, Deep Learning, Equine Kinematic Gait Analysis, Stereo Matching.

Effects of Image Quality on the Accuracy Human Pose Estimation and Detection of Eye Lid Opening/Closing Using Openpose and DLib.

The application of computer models in continuous patient activity monitoring using video cameras is complicated by the capture of images of varying qualities due to poor lighting conditions and lower image resolutions. Insufficient literature has assessed the effects of image resolution, color depth, noise level, and low light on the inference of eye opening and closing and body landmarks from digital images. This study systematically assessed the effects of varying image resolutions (from 100 × 100 pixels to 20 × 20 pixels at an interval of 10 pixels), lighting conditions (from 42 to 2 lux with an interval of 2 lux), color-depths (from 16.7 M colors to 8 M, 1 M, 512 K, 216 K, 64 K, 8 K, 1 K, 729, 512, 343, 216, 125, 64, 27, and 8 colors), and noise levels on the accuracy and model performance in eye dimension estimation and body keypoint localization using the Dlib library and OpenPose with images from the Closed Eyes in the Wild and the COCO datasets, as well as photographs of the face captured at different light intensities. The model accuracy and rate of model failure remained acceptable at an image resolution of 60 × 60 pixels, a color depth of 343 colors, a light intensity of 14 lux, and a Gaussian noise level of 4% (i.e., 4% of pixels replaced by Gaussian noise). The Dlib and OpenPose models failed to detect eye dimensions and body keypoints only at low image resolutions, lighting conditions, and color depths. Our established baseline threshold values will be useful for future work in the application of computer vision in continuous patient monitoring.

A Real-Time Remote Respiration Measurement Method with Improved Robustness Based on a CNN Model

Human respiration reflects meaningful information, such as one’s health and psychological state. Rates of respiration are an important indicator in medicine because they are directly related to life, death, and the onset of a serious disease. In this study, we propose a noncontact method to measure respiration. Our proposed approach uses a standard RGB camera and does not require any special equipment. Measurement is performed automatically by detecting body landmarks to identify regions of interest (RoIs). We adopt a learning model trained to measure motion and respiration by analyzing movement from RoI images for high robustness to background noise. We collected a remote respiration measurement dataset to train the proposed method and compared its measurement performance with that of representative existing methods. Experimentally, the proposed method showed a performance similar to that of existing methods in a stable environment with restricted motion. However, its performance was significantly improved compared to existing methods owing to its robustness to motion noise. In an environment with partial occlusion and small body movement, the error of the existing methods was 4–8 bpm, whereas the error of our proposed method was around 0.1 bpm. In addition, by measuring the time required to perform each step of the respiration measurement process, we confirmed that the proposed method can be implemented in real time at over 30 FPS using only a standard CPU. Since the proposed approach shows state-of-the-art accuracy with the error of 0.1 bpm in the wild, it can be expanded to various applications, such as medicine, home healthcare, emotional marketing, forensic investigation, and fitness in future research.

Automatic extraction of upper body landmarks using Rhino and Grasshopper algorithms

The aim of this study is to develop algorithms for automatic landmark extraction on women with various upper body types and body inclinations using the Grasshopper algorithm editor, which enables the user to interact with the 3D modeling interface. First, 15 landmarks were defined based on the morphological features of 3D body surfaces and clothing applications, from which automatic landmark extraction algorithms were developed. To verify the accuracy of the algorithms on various body shapes, this study determined criteria for key body shape factors (BMI, neck slope, upper body slope, and shoulder slope) that influence each landmark position, classified them into body shape groups and sorted the scan samples for each body type using the 6th SizeKorea database. The statistical differences between the scan-derived measurements and the SizeKorea measurements were compared, with an allowable tolerance of ISO 20685. In the case of landmarks with significant differences, the algorithm was modified. It was found that the algorithms were successfully applied to various upper body shapes, which improved the reliability and accuracy of the algorithm.

OpenMonkeyChallenge: Dataset and Benchmark Challenges for Pose Estimation of Non-human Primates.

The ability to automatically estimate the pose of non-human primates as they move through the world is important for several subfields in biology and biomedicine. Inspired by the recent success of computer vision models enabled by benchmark challenges (e.g., object detection), we propose a new benchmark challenge called OpenMonkeyChallenge that facilitates collective community efforts through an annual competition to build generalizable non-human primate pose estimation models. To host the benchmark challenge, we provide a new public dataset consisting of 111,529 annotated (17 body landmarks) photographs of non-human primates in naturalistic contexts obtained from various sources including the Internet, three National Primate Research Centers, and the Minnesota Zoo. Such annotated datasets will be used for the training and testing datasets to develop generalizable models with standardized evaluation metrics. We demonstrate the effectiveness of our dataset quantitatively by comparing it with existing datasets based on seven state-of-the-art pose estimation models.

High-Resolution Three-Dimensional Hybrid MRI + Low Dose CT Vocal Tract Modeling: A Cadaveric Pilot Study

<h2>Summary</h2><h3>Objectives</h3> MRI based vocal tract models have many applications in voice research and education. These models do not adequately capture bony structures (e.g. teeth, mandible), and spatial resolution is often relatively low in order to minimize scanning time. Most MRI sequences achieve 3D vocal tract coverage at gross resolutions of 2 mm³ within a scan time of <20 seconds. Computed tomography (CT) is well suited for vocal tract imaging, but is infrequently used due to the risk of ionizing radiation. In this cadaveric study, a single, extremely low-dose CT scan of the bony structures is blended with accelerated high-resolution (1 mm³) MRI scans of the soft tissues, creating a high-resolution hybrid CT-MRI vocal tract model. <h3>Methods</h3> Minimum CT dosages were determined and a custom 16-channel airway receiver coil for accelerated high (1 mm³) resolution MRI was evaluated. A rigid body landmark based partial volume registration scheme was then applied to the images, creating a hybrid CT-MRI model that was segmented in Slicer. <h3>Results</h3> Ultra-low dose CT produced images with sufficient quality to clearly visualize the bone, and exposed the cadaver to 0.06 mSv. This is comparable to atmospheric exposures during a round trip transatlantic flight. The custom 16-channel vocal tract coil produced acceptable image quality at 1 mm³ resolution when reconstructed from ∼6 fold undersampled data. High (1 mm³) resolution MR imaging of short (<10 seconds) sustained sounds was achieved. The feasibility of hybrid CT-MRI vocal tract modeling was successfully demonstrated using the rigid body landmark based partial volume registration scheme. Segmentations of CT and hybrid CT-MRI images provided more detailed 3D representations of the vocal tract than 2 mm³ MRI based segmentations. <h3>Conclusions</h3> The method described in this study indicates that high-resolution CT and MR image sets can be combined so that structures such as teeth and bone are accurately represented in vocal tract reconstructions. Such scans will aid learning and deepen understanding of anatomical features that relate to voice production, as well as furthering knowledge of the static and dynamic functioning of individual structures relating to voice production.

Development of a multi-modal sensor network to detect and monitor knee joint condition

Osteoarthritis is a major cause of mobility problems in older people and is a particular problem in former sportspeople. The objective of this study was to develop and characterise a new system for the detection, monitoring and analysis of acoustic emissions from knee joints. 15 adult volunteers participated in the study. The participants performed six sets of three sit-stand-sit cycles. Reflective markers were placed at specific body landmarks recorded by 3D cameras. The exercise was performed with one foot on a force platform. A sensitive condenser microphone with a wide frequency response was connected to a dedicated acoustic analysis unit. Preliminary results provide clear acoustic signals showing a distinctive sequence of impulse-decay forms occurring naturally during each sit-stand-sit cycle. There are distinct differences between the acoustic signals emitted from younger healthy knees and those from aged knees. This work demonstrates the potential for this system to be used as an indication of the state of health of a human knee during movement.

Criterion Validity Of Tri-monitor Activpal Configuration In Determining Knee-flexion Angles During Sitting In Laboratory Setting

Prolonged sitting impairs cardiovascular and metabolic health. However, studies in this area have been limited to the laboratory setting, primarily due to the lack of measurement tools to characterize habitual sedentary behaviours. Therefore, quantifying the amount of time that people sit in a knee bent posture, causing arterial kink, may help determine if lower limb sedentary postures play a role in cardiovascular health. ActivPAL sensors located on the thigh and torso have been shown to successfully distinguish sitting from laying postures. An additional activPAL monitor positioned on the shin may be able to determine knee flexion angles and further characterize free-living sedentary behaviours. PURPOSE: Test the hypothesis that a tri-monitor activPAL setup can accurately measure knee flexion angles during sitting compared to motion capture (criterion). METHODS: Nineteen adults (12 ♀, 24 ± 4 years, 23.2 ± 4.0 kg/m2) wore three activPAL monitors (torso, thigh, shin) while 14 motion capture cameras simultaneously collected 15 markers located on body landmarks around the ankle, knee, hip, and shoulders. Each participant completed eight 45-s seated trials followed by 15-s of standing. The trial knee flexion angles (105°, 90°, 75°, 60°, 45°, 30°, 15°, 0°) were determined using a handheld goniometer. Validity was assessed via repeated measures ANOVA, Bland-Altman analyses, and intraclass correlations. RESULTS: Compared to motion capture, the calculated angles from activPALs were not different across 15-75° angles. The activPAL underestimated knee flexion angles at 90-105° (difference: ~5.0°; all, p ≤ 0.02), but overestimated knee flexion angles at 0° (difference: 9.1°, p < 0.001). A fixed bias (-0.3 ± 7.9°, p = 0.665) was not observed but a positive proportional bias was (β = 0.223, p < 0.001). The activPAL angles were highly correlated to those from the motion capture (ICC = 0.98, p < 0.001). CONCLUSION: The activPAL is an accurate measure to determine seated knee flexion angles from 75° to 15° but exhibits an average 5–10° error at knee flexion angles >75° and < 15°. Although there is error at the end ranges of motion, the tri-monitor activPAL configuration may be a useful tool to quantify sedentary knee flexion patterns during free-living conditions.

Simple method integrating OpenPose and RGB-D camera for identifying 3D body landmark locations in various postures

Ergonomic assessments of posture are indispensable for reducing the risk of physical discomfort in the workplace. However, it is challenging to measure postural data in field studies because of the high cost and complex setup required by conventional motion tracking systems. OpenPose is an advanced approach for real-time multi-person two-dimensional (2D) pose estimations in an image, whereas RGB-D cameras can simultaneously record image and depth data. Thus, the present study proposes integrating these two approaches to identify 3D body landmark locations. To quantify the accuracy of the proposed method, the anatomical body landmark locations identified by a marker-based reference system were used as a gold standard. The tracking errors of using two RGB-D cameras, which used different data acquisition techniques (stereoscopic and time-of-flight (ToF)), were examined and compared. Thirty participants were recruited to perform 15 static postures. The average tracking errors of the landmark locations were 7.96 ± 3.59 and 9.81 ± 5.57 cm (stereoscopic), and 6.38 ± 2.88 and 8.18 ± 5.56 cm (ToF) for the standing postures and sitting postures, respectively. Depending on the desired accuracy, the integration of the RGB-D cameras and OpenPose could provide an alternative motion tracking method to identify 3D body landmark locations in a simple manner when postural assessments are performed.

This video is going to show you how it works to identify the target nodule by using positioning method of 3D localization based on anatomical landmarks. A 71 years old male patient with a mixed ground glass nodule was found in the lower lobe of left lung. First, the head of the transverse process of the vertebral body as point

This video is going to show you how it works to identify the target nodule by using positioning method of 3D localization based on anatomical landmarks. A 71 years old male patient with a mixed ground glass nodule was found in the lower lobe of left lung. First, the head of the transverse process of the vertebral body as point

Body landmark detection with an extremely small dataset using transfer learning

Body landmark detection with an extremely small dataset using transfer learning

Postural analysis of side impact WorldSID 50th ATD in highly reclined seats

Objective The anthropomorphic test device (ATD) used in NCAP frontal impact, the Thor 50th, has undergone substantial changes for safety restraint systems testing of upcoming autonomous vehicles (AVs), resulting in the Thor-AV 50 M. Likewise, a deep analysis of the ATD for lateral impacts, the WorldSID 50th, is needed to determine the requirements it has to fulfill to accurately replicate the human body response under lateral impact. This study sets out general considerations regarding the WorldSID 50th ATD design in order to assess its suitability for AV testing scenarios. Methods The WorldSID 50th ATD was placed in a driver seat at a conventional upright driving posture. Relevant body landmarks and angles as well as profile streams of the mockup and the ATD were recorded. The seatback was then reclined stepwise. The ATD was repositioned and the recording procedure repeated at every posture until the seatback was fully reclined. Examination of the ATD assemblies and hinge mechanisms was performed, searching for limitations that could result on inadequate positioning or performance under impact. An eventual SAEJ3016 Level 3 cross-legged posture (SAE J3016, 2021) was considered for examination as well. Results Most relevant anthropometric constraints were noticed at the abdomen and the shoulder and neck assemblies: At the abdomen, a gap between the pelvis and the lower torso appeared at seatback recline angles larger than 30°, likely promoting belt intrusion in case of a test event, and the lumbar spine rubber deformed progressively in a manner that might cause tearing damage. The number and arrangement of the shoulder clevis detents, on the other hand, did not allow the arm to be placed at intermediate positions. Finally, the current neck bracket design allowed head orientation only along the sagittal plane. The head could neither be level at seatback angles larger than 43° nor be placed resting on the head restraint for seatback angle values larger than 59°. Conclusions The WorldSID ATD stands as an accurate 50th percentile human being surrogate in upright and low/mid reclined postures. Some limitations that might constrain its use in AV scenarios, though, appeared at mid/high reclined positions. Hence, design updating of the lumbar, neck and shoulder clevis assemblies, and the implementation of instrumented lower abdomen and lower arms assemblies should be taken into consideration.

Emotion and Gesture detection

Abstract: Machine learning algorithms have removed the constraints of computer vision. Researchers and developers have developed new approaches for detecting emotions making it possible to predict the behaviour and consecutive actions of human beings. As machine learning methods make use of GPUs' massive computation capability, these models' image processing skills are well suited to real-world issues. Computer vision has moved from a niche field to a variety of other fields, including behavioural sciences. These algorithms or models are utilised in a wide range of real-world applications, including security, driver safety, autonomous cars, human-computer interaction, and healthcare. Due to the emergence of graphics processing units, which are hardware devices capable of doing millions of computations in seconds or minutes, these models are constantly changing. Technologies like augmented reality and virtual reality are also on the rise. Robotic vision and interactive robotic communication are two of their most intriguing uses. Both verbal and visual modalities can be used to identify human emotions. Facial expressions are an excellent way to determine a person's emotional state. This work describes a real-time strategy for emotion and gesture detection. The fundamental idea is to use the MediaPipe framework which is based on real-time deep learning, to generate critical points. Furthermore, A series of precisely constructed mesh generators and angular encoding modules are used to encode the generated key points. Finally, by assessing failure instances of existing models, we are evaluating the applicability of emotion and gesture detection from our model.We are using models such as Random Forest(RF),logistic regression(LR),Gradient Classifier(GR) and Ridge classifier(RC). Real-time inference and good prediction quality are demonstrated by the suggested system and architecture. Keywords: Body Landmarks, MediaPipe, Prediction, Accuracy, Real-time on-device Tracking, Recognition.

Camera fusion for real-time temperature monitoring of neonates using deep learning

The continuous monitoring of vital signs is a crucial aspect of medical care in neonatal intensive care units. Since cable-based sensors pose a potential risk for the immature skin of preterm infants, unobtrusive monitoring techniques using camera systems are increasingly investigated. The combination of deep learning–based algorithms and camera modalities such as RGB and infrared thermography can improve the development of cable-free methods for the extraction of vital parameters. In this study, a real-time approach for local extraction of temperatures on the body surface of neonates using a multi-modal clinical dataset was implemented. Therefore, a trained deep learning–based keypoint detector was used for body landmark prediction in RGB. Image registration was conducted to transfer the RGB points to the corresponding thermographic recordings. These landmarks were used to extract the body surface temperature in various regions to determine the central-peripheral temperature difference. A validation of the keypoint detector showed a mean average precision of 0.82. The registration resulted in mean absolute errors of 16.4 px (8.2 mm) for x and 22.4 px (11.2 mm) for y. The evaluation of the temperature extraction revealed a mean absolute error of 0.55 ^{circ }C. A final performance of 31 fps was observed on the NVIDIA Jetson Xavier NX module, which proves real-time capability on an embedded GPU system. As a result, the approach can perform real-time temperature extraction on a low-cost GPU module.Graphical abstract

The vital landmarks of human body (Marma) according to TCAM and its application in Physical medicine and Rehabilitation

According to Traditional Complementary and Alternative Medicine (TCAM), the universe as well as the human body, is grossly and subtly comprised of five elements. The five elements namely space, air, fire, water and earth. These are the prime factors for origin of life. Among the five elements, space is related with mind, air with body movements, fire with thermodynamics of the body, water with structural integrity of the body and earth with gross body. The Indian System of Medical science has contributed various insights into the understanding of human body using various philosophical, analytical, experimental and spiritual modalities. Marma sastra and varma sastraare two nomenclatures given for the same type of human anatomical body landmarks, which strongly works on the principles of anatomy and physiology and energy principles. Varma/marma is air (vayu). Varma is also referred to as energy. Thirteen energy locations, grouped into three categories has been explained in detail. In order to understand the varma techniques, an additional knowledge regarding the basic Indian philosophy, is also needed.Experts have given various classifications of marma. Among them the most practical system is based on method of application. It illustrates the following: 1) Thodu Varmam (stimulated by touching), 2) Padu Varmam (activate other regions), 3) Thattu Varmam (stimulated by gentle tap), 4) Enai Varmam (unified activation of two points can relieve symptoms), 5) Pakka Varmam (action of adjacent varmas), 6) Ellidai Varmam (points related in between joints), 7) Oothu Varmam (stimulation by blowing; for children), 8) Nakku Varmam (stimulation by licking; for children), 9) Naal Varmam (associated with astrological stars), 10) Griha Varma (associated with astrological planets), 11) Patchi Varmam (points related to internal organs), 12) Vaatha Varmam, 13) Pitha Varmam, 14) Sleshma Varmam, (12, 13 and 14 areregulated by bodily humor 15) Nokku Varmam (stimulated by looking). More than fifty manuscripts have been published in the Nation Workshop on Manuscriptology, India, 1999. The present session is a representation of the major varma/marma landmarks of human body and its utility based on connecting this physical landmarks to improve the mobility standards of human body. These principles will be detailed using most recent photographs from eminent resources followed by detailed diagrammatic representation of their practical applications. Though based on ancient knowledge, the matter detailed in this session is based on current practice of these methods in India. The topic gives a new dimension and scope in using the Varma/Marma as a anatomico‐physiological landmarks for medical students as well as practitioners of physical medicine and rehabilitation (PMR).