Recent advances in error rate estimation

Different types of error rates are described and the state of the art of error rate estimation at the time of Toussaint's (1974) survey is briefly summarised. Developments since then are outlines, and the two major advances, namely bootstrap and average conditional error rate estimation methods, and their extensions, are described in detail.     

Oxygen consumption during outdoor recreational cycling

The relation between oxygen consumption and cycling speed during outdoor recreational cycling was studied in 20 healthy subjects, men and women aged 20–30 and 50–60 years. They rode a touring bicycle at speeds between 2·8 and 8·3ms^?1(10 and 30km h^?1). No significant differences in oxygen consumption were found between the sexes and two age groups. The scatter of the oxygen consumption data was least when oxygen consumption was expressed in terms of body surface. Different types of equations were developed for the prediction of oxygen consumption from the cycling speed which all gave approximately equally accurate predictions ( is oxygen consumption in lmin^?1 m^?2, Vis, cycling speed in m s^?1). The equation may also be applied for cycling in traffic situations if the mean speed is corrected for stop limes.     

Effect of wearing personal protective clothing and self-contained breathing apparatus on heart rate, temperature and oxygen consumption during stepping exercise and live fire training exercises

Fire fighter breathing apparatus instructors (BAIs) must possess the ability to respond to both the extrinsic stress of a high temperature environment and the intrinsic stress from wearing personal protective equipment (PPE) and self-contained breathing apparatus (SCBA), repeatedly and regularly, whilst training recruits in live fire training exercises (LFTEs). There are few previous investigations on BAIs in hot environments such as LFTEs, since the main research focus has been on regular fire fighters undertaking exercises in temperate or fire conditions at a moderate to high exercise intensity. In this study, the intrinsic cardiovascular stress effects of wearing PPE + SCBA were first investigated using a step test whilst wearing gym kit (control), weighted gym kit (a rucksack weighted to the equivalent of PPE + SCBA) and full PPE + SCBA (weight plus the effects of protective clothing). The extrinsic effects of the very hot environment were investigated in BIAs in LFTEs compared to mock fire training exercises (MFTEs), where the fire was not ignited. There was an increase in heart rate due to the modest workload imposed on the BAIs through carrying out the MFTEs (25.0 (18.7)%) compared to resting. However, when exposed to fire during the LFTEs, heat storage appears to be significant as the heart rate increased by up to 39.8 (+/-20.1)% over that of the mock LFTEs at temperate conditions. Thus, being able to dissipate heat from the PPE is particularly important in reducing the cardiovascular responses for BAIs during LFTEs.     

Effect of wearing personal protective clothing and self-contained breathing apparatus on heart rate,temperature and oxygen consumption during stepping exercise and live fire training exercises

Fire fighter breathing apparatus instructors (BAIs) must possess the ability to respond to both the extrinsic stress of a high temperature environment and the intrinsic stress from wearing personal protective equipment (PPE) and self-contained breathing apparatus (SCBA), repeatedly and regularly, whilst training recruits in live fire training exercises (LFTEs). There are few previous investigations on BAIs in hot environments such as LFTEs, since the main research focus has been on regular fire fighters undertaking exercises in temperate or fire conditions at a moderate to high exercise intensity. In this study, the intrinsic cardiovascular stress effects of wearing PPE + SCBA were first investigated using a step test whilst wearing gym kit (control), weighted gym kit (a rucksack weighted to the equivalent of PPE + SCBA) and full PPE + SCBA (weight plus the effects of protective clothing). The extrinsic effects of the very hot environment were investigated in BIAs in LFTEs compared to mock fire training exercises (MFTEs), where the fire was not ignited. There was an increase in heart rate due to the modest workload imposed on the BAIs through carrying out the MFTEs (25.0 (18.7)%) compared to resting. However, when exposed to fire during the LFTEs, heat storage appears to be significant as the heart rate increased by up to 39.8 (±20.1)% over that of the mock LFTEs at temperate conditions. Thus, being able to dissipate heat from the PPE is particularly important in reducing the cardiovascular responses for BAIs during LFTEs.     

Fighting fire and fatigue: sleep quantity and quality during multi-day wildfire suppression

This study examined firefighters’ sleep quantity and quality throughout multi-day wildfire suppression, and assessed the impact of sleep location, shift length, shift start time and incident severity on these variables. For 4 weeks, 40 volunteer firefighters’ sleep was assessed using wrist actigraphy. Analyses revealed that the quantity of sleep obtained on fire days was restricted, and pre- and post-sleep fatigue ratings were higher, compared to non-fire days. On fire days, total sleep time was less when: (i) sleep location was in a tent or vehicle, (ii) shifts were greater than 14 h and (iii) shifts started between 05:00 and 06:00 h. This is the first empirical investigation providing objective evidence that firefighters’ sleep is restricted during wildfire suppression. Furthermore, sleep location, shift length and shift start time should be targeted when designing appropriate controls to manage fatigue-related risk and preserve firefighters’ health and safety during wildfire events.

Practitioner Summary: During multi-day wildfire suppression, firefighters’ sleep quantity was restricted, and pre- and post-sleep fatigue ratings were higher, compared to non-fire days. Furthermore, total sleep time was less when: (i) sleep occurred in a tent/vehicle, (ii) shifts were >14 h and (iii) shifts started between 05:00 and 06:00 h.     

A new heart rate variability-based method for the estimation of oxygen consumption without individual laboratory calibration: application example on postal workers

Traditionally, the estimation of oxygen consumption (VO2) at work using heart rate (HR) has required the determination of individual HR/VO2 calibration curves in a separate exercise test in a laboratory (VO2-TRAD). Recently, a new neural network-, and heart rate variability-based method has been developed (Firstbeat PRO heartbeat analysis software) for the estimation of VO2 without individual calibration (VO2-HRV). In the present study, the VO2-values by the VO2-HRV were compared with the values by VO2-TRAD in 22 postal workers. Within individuals the correlation between the two methods was high (range 0.80-0.99). The VO2-TRAD gave higher values of VO2 compared to VO2-HRV (19%) especially during low physical activity work when non-metabolic factors may increase HR. When assessed in different HR categories, the smallest difference (11%), and highest correlations (range 0.83-0.99) in VO2 between the methods were observed at higher HR levels. The results indicate that the VO2-HRV is a potentially useful method to estimate VO2 in the field without laboratory calibration.     

Re-sampling procedures for reducing bias of error rate estimation in multinomial classification

The bias in estimating the error rate of a multinomial classification problem is addressed. The cross-validation, jackknife and bootstrap procedures are introduced to obtain the bias-reducing estimates of error rate. The performance of these procedures is examined by Monte Carlo simulation under different population structures. It is found that all three procedures are effective in reducing bias under small sample situations.     

融合多头自注意力机制的无接触心率估计模型

为了在光照变化和头部运动条件下实现准确稳定的无接触心率估计,基于U-Net模型提出一种融合多头自注意力机制的端到端心率估计模型rPPG-UNet。该模型通过使用U型编码器—解码器网络结构实现对生理特征的提取与重建,并使用Skip Connection连接编码器与解码器实现浅层时间特征的复用。该模型还融合多头自注意力机制来捕获生理特征的时间依赖性。最后,该模型采用多任务学习策略以提高心率估计的准确度,加速网络训练。在公开数据集上的实验结果表明,rPPG-UNet的性能优于其他基线模型,可以实现更准确的无接触心率估计。     

Spatio-temporal filtering of thermal video sequences for heart rate estimation

In this paper, a novel method for non-contact measurement of heart rate using thermal imaging was proposed. Thermal videos were recorded from subjects’ faces. The measurements are performed on three different areas: the whole face, the upper half of the face and the supraorbital region. A tracker was used to track these regions to make the algorithm invulnerable to the subject's motion. After tracking, the videos were spatially filtered using a full Laplacian pyramid decomposition to increase the signal to noise ratio; next, the video frames were successively temporally filtered using an ideal bandpass filter for extracting the thermal variations caused by blood circulation. Finally, the heart rate was calculated by using two methods including zero crossing and Fast Fourier Transform. For evaluating the results, the complement of absolute normalized difference (CAND) index was used which was introduced by Pavlidis. This index was 99.42% in the best case and 92.472% in average for 22 subjects. These results showed a growth in CAND index in comparison with previous work. Zerocrossing outperformed FFT because of the nonstationary nature of thermal signals. Another benefit of our method is that, the videos are taken from the face unlike most of the studies that take it from the neck and Carotid. Neck and carotid are less accessible than faces. Finally, the optimum ROI for estimating the heart rate from face was identified.     

Relationship between fire fighting suppression tasks and physical fitness.

This investigation evaluated the relationship between physical fitness and performance of fire suppression tasks. The following mean +/- SD values were associated with 91 fire fighters: age 31.69 +/- 7.39 years, height 177.29 +/- 6.38 cm, weight 83.97 +/- 10.86 kg, % fat 13.78 +/- 4.31, fat free weight (FFW) 71.52 +/- 7.66 kg, pull-ups 9.03 +/- 4.79, push-ups 41.02 +/- 14.08, 1.5 mile run 737.60 +/- 108.11 s, sit and reach 32.00 +/- 8.5 cm, sit-ups 39.88 +/- 7.75, and total grip strength 116.75 +/- 17.67 kg. The physical performance assessment (PPA) consisted of the following: stair climb, hoist, forcible entry, hose advance, and victim rescue. Significant correlations (p < 0.01) were found between the PPA and the following: total grip strength (r = -0.54), FFW (r = -0.47), height (r = -0.40), pull-ups (r = -0.38), push-ups (r = -0.38), 1.5 mile run (r = 0.38), sit-ups (r = -0.32), weight (r = -0.30) and % fat (r = 0.30). Multiple regression analysis indicated that the best multiple predictor of PPA was the 1.5 mile run, FFW, and pull-ups, r = 0.73, p < 0.001. This investigation shows the importance of physical fitness as related to performance of fire suppression job tasks.     

Mathematical model of heart rate regulation during exercise

A model of the heart rate regulation during physical load and recovery has been developed. The model contains four parameters with the following physiological meanings: Heart rate increment related to the load of 1 watt, a load which can be coped with by vagal inhibition, a time constant of the vagal component and a time constant of the slower complex neurohumoral sympathetic component. All these parameters were estimated uniquely with relatively small reliability intervals for a group of 15 healthy men. The results are useful practically for the evaluation of the physical efficiency of individuals and whole groups of healthy subjects.     

Fundamental matrix estimation: A study of error criteria

The fundamental matrix (FM) describes the geometric relations that exist between two images of the same scene. Different error criteria are used for estimating FMs from an input set of correspondences. In this paper, the accuracy and efficiency aspects of the different error criteria are studied. We mathematically and experimentally proved that the most popular error criterion, the symmetric epipolar distance, is biased. It was also shown that despite the similarity between the algebraic expressions of the symmetric epipolar distance and Sampson distance, they have different accuracy properties. In addition, a new error criterion, Kanatani distance, was proposed and proved to be the most effective for use during the outlier removal phase from accuracy and efficiency perspectives. To thoroughly test the accuracy of the different error criteria, we proposed a randomized algorithm for Reprojection Error-based Correspondence Generation (RE-CG). As input, RE-CG takes an FM and a desired reprojection error value d. As output, RE-CG generates a random correspondence having that error value. Mathematical analysis of this algorithm revealed that the success probability for any given trial is 1 − (2/3)² at best and is 1 − (6/7)² at worst while experiments demonstrated that the algorithm often succeeds after only one trial.     

基于低秩稀疏矩阵分解的非接触心率估计

心率检测作为一项重要的生理检测指标,在医学健康、刑侦检测、信息安全等方 面具有重要应用。计算机视觉领域近期的研究表明,心率信号可以通过摄像头捕捉的视频予以 获取。现有的研究方法在理想的实验环境下已取得较好的效果,然而在自然状态面部旋转以及 出现各种噪声(阴影、遮挡)时鲁棒性较弱。通过检测人脸的关键点,获得面部区域的感兴趣, 避免因面部旋转引入检测误差,在现有模型的基础上提出一种基于低秩稀疏矩阵分解的非接触 式心率估计模型,对频域血液体积脉冲(BVP)信号矩阵实现去噪处理,解决使用摄像头非接触 式获取心率信号时存在的问题。实验显示,该模型在MAHNOB-HCI 数据集上实现了3.25%的 误差比均值,优于现有的模型。     

Bounded error parameter estimation: a sequential analytic centerapproach

A sequential analytic center approach for bounded error parameter estimation is proposed. The authors show that the analytic center minimizes the logarithmic average output error among all the estimates within the membership set and is a maximum likelihood estimator for a class of noise density functions which include parabolic densities and approximations of truncated Gaussian. They also show that the analytic center is easily computable for both offline and online problems with a sequential algorithm. The convergence proof of this sequential algorithm is obtained and, moreover, it is shown that the complexity in terms of the maximum number of Newton iterations is linear in the number of observed data points     

Effects of dynamic,static and combined work on heart rate and oxygen consumption

The effects upon heart rate and oxygen consumption of muscular exercises including simultaneous dynamic and static contractions were studied in three male subjects. Dynamic work consisted of walking at four speeds (0·56, 0·83, 1·11, 1·39 m s^?) on a horizontal treadmill; static work consisted of pushing against, pulling and holding 6, 9, 12, 18 and 24?kg; combined work associated walking with each one of the forms of static work. Physiological load is expressed in terms of cardiac cost (ΔHR) and oxygen cost (Δ[Vdot]_o2). The physiological cost of combined work increases with both the walking speed and the static load. For each parameter (HR and [Vdot]_o2) the extra-cost of combined work has been determined by computing the difference between the cost of combined work and the sum of the costs observed during static and dynamic exercises separately performed. The paired t-test shows significant differences for all of the walking-pushing tests, but only for 8 pulling tests and 2 holding tests. Linear relationships are observed between the oxygen extra-cost and load when walking at 0·56 or 0·83 ms^?1, with correlation coefficients statistically significant for pushing and pulling (p < 0·01) but not significant for holding tests. The present results show that, when the static work is combined with walking, the physiological response varies with the type of static work considered.     

Oxygen consumption,heat production,and muscular efficiency during uphill and downhill walking

Oxygen consumption, heat production, and muscular efficiency for walking are parameters important to know for ergonomics models and equipment design. Most of these assume that the oxygen consumption and heat production of downhill walking are the same as for uphill walking. Eight subjects wearing insulating clothing walked on a treadmill at three uphill and three downhill grades, and at level grade at a rate of 1.1 m/s. Oxygen consumption VO2 was calculated from steady state measurements of respiratory minute volume and oxygen percentage. Heat production (q) was calculated from the rate of heat storage in the body and clothing. Least-squares best fit equations for oxygen consumption and heat production found were to be VO2 = 0.813 + 0.0361G + 0.000810G2 - 0.0000302G3 and q = 6.55 + 0.185G + 0.0114G2 - 0.000190G3 where G is percent grade. Approximations show that VO2 (downhill) = 0.5 VO2 (uphill), q (downhill) = 0.67 q (uphill), and muscular efficiency eta (downhill) = -2eta (uphill).