A method for improving semantic segmentation using thermographic images in infants

Background Regulation of temperature is clinically important in the care of neonates because it has a significant impact on prognosis. Although probes that make contact with the skin are widely used to monitor temperature and provide spot central and peripheral temperature information, they do not provide details of the temperature distribution around the body. Although it is possible to obtain detailed temperature distributions using multiple probes, this is not clinically practical. Thermographic techniques have been reported for measurement of temperature distribution in infants. However, as these methods require manual selection of the regions of interest (ROIs), they are not suitable for introduction into clinical settings in hospitals. Here, we describe a method for segmentation of thermal images that enables continuous quantitative contactless monitoring of the temperature distribution over the whole body of neonates. Methods The semantic segmentation method, U-Net, was applied to thermal images of infants. The optimal combination of Weight Normalization, Group Normalization, and Flexible Rectified Linear Unit (FReLU) was evaluated. U-Net Generative Adversarial Network (U-Net GAN) was applied to thermal images, and a Self-Attention (SA) module was finally applied to U-Net GAN (U-Net GAN + SA) to improve precision. The semantic segmentation performance of these methods was evaluated. Results The optimal semantic segmentation performance was obtained with application of FReLU and Group Normalization to U-Net, showing accuracy of 92.9% and Mean Intersection over Union (mIoU) of 64.5%. U-Net GAN improved the performance, yielding accuracy of 93.3% and mIoU of 66.9%, and U-Net GAN + SA showed further improvement with accuracy of 93.5% and mIoU of 70.4%. Conclusions FReLU and Group Normalization are appropriate semantic segmentation methods for application to neonatal thermal images. U-Net GAN and U-Net GAN + SA significantly improved the mIoU of segmentation.

Passive detection of COVID-19 with wearable sensors and explainable machine learning algorithms

Individual smartwatch or fitness band sensor data in the setting of COVID-19 has shown promise to identify symptomatic and pre-symptomatic infection or the need for hospitalization, correlations between peripheral temperature and self-reported fever, and an association between changes in heart-rate-variability and infection. In our study, a total of 38,911 individuals (61% female, 15% over 65) have been enrolled between March 25, 2020 and April 3, 2021, with 1118 reported testing positive and 7032 negative for COVID-19 by nasopharyngeal PCR swab test. We propose an explainable gradient boosting prediction model based on decision trees for the detection of COVID-19 infection that can adapt to the absence of self-reported symptoms and to the available sensor data, and that can explain the importance of each feature and the post-test-behavior for the individuals. We tested it in a cohort of symptomatic individuals who exhibited an AUC of 0.83 [0.81–0.85], or AUC = 0.78 [0.75–0.80] when considering only data before the test date, outperforming state-of-the-art algorithm in these conditions. The analysis of all individuals (including asymptomatic and pre-symptomatic) when self-reported symptoms were excluded provided an AUC of 0.78 [0.76–0.79], or AUC of 0.70 [0.69–0.72] when considering only data before the test date. Extending the use of predictive algorithms for detection of COVID-19 infection based only on passively monitored data from any device, we showed that it is possible to scale up this platform and apply the algorithm in other settings where self-reported symptoms can not be collected.

Gaps in neonatal thermal care in low-resource settings: A web-based survey of healthcare workers.

Objective To explore the gaps in knowledge and practice in neonatal thermal care among healthcare workers in low-resource settings. Methods We conducted a 2-round, web-based survey of a purposive and snowball sample of healthcare workers in neonatal care in low-resource settings globally. The questionnaire was developed using themes of neonatal thermal care extracted from existing neonatal care guidelines, including WHO’s. The survey asked multiple-choice questions, supplemented by open-ended questions to capture first hand insights and information on neonatal thermal care. Results of the survey were analysed using Microsoft Excel. Data was collated and summarized using descriptive measures. Results Almost all participants acknowledged the importance of all the WHO warm chain elements, however, fewer participants responded positively regarding the practice of this warm chain. Only 56% of the participants acknowledged the usefulness of checking the peripheral temperature by hand-touch. The usefulness of the core temperature was valued higher than that of the peripheral temperature as an indicator of cold stress, with 70% and 58% agreement, respectively. Opinions diverged regarding the peripheral temperatures, including apparent inaccuracy compared to rectal or axillary temperature. Preferences on rewarming strategies widely differed among participants and so did the availability of warming equipment in their institutions. Conclusions We identified the general acknowledgment of the importance of the WHO warm chain but also its limited practice. We also identified that an inadequate understanding of cold stress underestimates the potential benefits of peripheral temperatures and leads to missed opportunities for timely prevention of hypothermia. Furthermore, lack of consistent guidance on equipment for rewarming hypothermic neonates hampers recovery.

Does Cues on Mother’s Face Convey the Condition of the Child in a Critical Care Unit - A Prospective Cross-Sectional Study Done in the Paediatric Intensive Care Unit, MGM Hospital, Warangal

BACKGROUND Face is the index of mind. All human beings tend to express various emotions through different facial expressions. Parents of children who are critically ill, experience both positive and negative emotions simultaneously as they may feel love, affection and sometimes sad, angry, and helpless. METHODS We observed facial expressions of mothers of children and recorded vital data of these children in our Paediatric Intensive Care Unit (PICU). 107 children were observed, and vitals recorded simultaneously by two different observers from time of admission to 72 hours at fixed time intervals, using Hamilton’s scoring system. RESULTS 107 critically ill children were observed. Glasgow Coma Scale (GCS), AVPU scale, peripheral temperature, capillary filling time (CFT), seizures at admission, were corelated with facial expressions of the mother at the time of admission, at 12 hours, at 24 hours, at 48 hours and at 72 hours. We found that facial expressions of the mothers changed from fearful/sad to happy as the vital data of the child improved with time interval. CONCLUSIONS At ‘0’ hour of admission, irrespective of the categorization of the GCS score, AVPU scores, peripheral temperature, and capillary filling time, 78 mothers were fearful, and 29 mothers were sad, amounting to a total of 107. None were happy, irrespective of the child’s clinical condition, as, the child being admitted, is in itself a stressful condition. There was a good increase in number of happy mothers in co-relation with child’s improvement. Hence by observing the mother’s face, we can definitely predict the clinical condition of the child and expressions on the mother’s face is a reflection of the clinical condition of the child. KEYWORDS Cues, Mother’s Face, Critically Ill Children

Passive Detection of COVID-19 with Wearable Sensors and Explainable Machine Learning Algorithms

Individual smartwatch or fitness band sensor data in the setting of COVID-19 has shown promise to identify symptomatic and pre-symptomatic infection or the need for hospitalization, correlations between peripheral temperature and self-reported fever, and an association between changes in heart-rate-variability and infection. In our study, a total of 38,911 individuals (61% female, 15% over 65) have been enrolled between March 25, 2020 and April 3, 2021, with 1,118 reported testing positive and 7,032 negative for COVID-19 by nasopharyngeal PCR swab test. We propose an explainable gradient boosting prediction model based on decision trees for the detection of COVID-19 infection that can adapt to the absence of self-reported symptoms and to the available sensor data, and that can explain the importance of each feature and the post-test-behavior for the individuals. We tested it in a cohort of symptomatic individuals who exhibited an AUC of 0.83 [0.81-0.85], or AUC=0.78 [0.75-0.80] when considering only data before the test date, outperforming state-of-the-art algorithm in these conditions. The analysis of all individuals (including asymptomatic and pre-symptomatic) when self-reported symptoms were excluded provided an AUC of 0.78 [0.76-0.79], or AUC of 0.70 [0.69-0.72] when considering only data before the test date. Extending the use of predictive algorithms for detection of COVID-19 infection based only on passively monitored data from any device, we showed that it is possible to scale up this platform and apply the algorithm in other settings where self-reported symptoms can not be collected.

Evaluation of systemic vasoconstriction and prognosis using thermography: a systematic review

Purpose: This systematic review investigates whether infrared thermography (IRT) can measure systemic vasoconstriction and addresses the value of IRT in assessing circulatory deficiency and prognoses. Methods: Design was based on the PRISMA criteria and a systematic search of 6 databases was performed. Results: Of 3,198 records, five articles were included. Three clinical studies were identified; two found significant correlations between IRT obtained temperatures and mortality. An experimental study found an association between peripheral temperature and stroke volume. An animal study found that central-peripheral temperature differences correlated with shock index, mean arterial pressure, and disease progression. Conclusions: Data from the most valid study suggests that central-peripheral temperature differences should be investigated further, both on its own, and integrated with other variables.

Thermographic visualization of facial vasoconstriction is associated with 30-day all-cause mortality in medical patients; prospective observational two-site cohort study

Introduction: Quick and reliable assessment of acute patients is required for accurate triage. The temperature gradient between core and peripheral temperature could possibly instantly provide information on circulatory status. Methods: Adult medical patients, who did not receive supplementary oxygen, attending two emergency departments, had a thermographic image taken on arrival. The association between 30-day mortality and gradients was tested using logistic regression. Results: 726 patients were studied, median age was 64 years and 14 (1.9%) died within 30 days. There was a significant association between mortality and temperature gradient, comparable to vital signs, age, and clinical intuition. Conclusion: Temperature gradient between nose and eye had an acceptable discriminatory power for 30-day all-cause mortality.

Pain Caused by Ear Tagging in Kids of Native Black Goats

There is no study about changes in behaviour and peripheral temperature of kid goats undergoing ear tagging procedure. This study was designed to elucidate that pain caused by ear tagging affects peripheral temperatures and behavioural observations in Karadi kid goats. Nineteen native black goat kids, aged 10 – 14 days, were used in this study. Eye and nasal temperatures were recorded before the ear tagging process, and after ear tagging 5 periods in 30 minutes. Besides, the researchers observed each kid's behaviour for 30 minutes before ear tagging and 30 minutes after it, using focal sampling method and the data recorded with instantaneous time sampling to measure the duration and frequency of each behaviour of the kids. Results revealed that peripheral temperatures were significantly decreased for both eye (P≤0.01) and ear (P≤0.01) after ear tagging. The temperatures of both eye and ear decreased after tagging significantly (P≤0.01) for 30 minutes. The proportion of time kid goats spent normal standing and suckling the dam’s teat decreased whereas the proportion of time spent head-shaking increased after tagging. Significant differences were found between abnormal standing (P≤0.01) and vocalization (P≤0.01). It is concluded that ear tagging causes a considerable pain in kid goats and using surface temperatures and behaviour are useful indicators to approve it.