基于高斯过程模型的指节图像识别方法

在基于图像的人机交互智能装配的手势识别与动作跟踪中,手部关节的图像定位是基础,并且关节信息的准确性对手势描述和行为识别与理解有直接影响.针对指节图像特征分布具有较强随机性,利用同态滤波进行图像预处理,以增强图像特征.基于高斯过程模型对手部指节图像二类特征进行学习,用样本对象的聚类测度,学习数据分布的特征模型,将学习获得的两类特征模型作为图像特征的检测器,检测结果即为图像的两个似然值.将经过正负类样本标记过的两种模型似然值作为输入,直接依据估计结果对手部关节图像进行检测识别.通过对不同位置处的手部关节识别分析和测试库检测,结果表明,本文所述方法可以直接得到后验概率的分布,提高了目标识别的准确性和效率.     

基于Laplace逼近Gaussian过程的指节图像中层偏移测度特征学习

在人机协调装配中,为了准确描述手部位姿,需要精确的指节图像特征提取与识 别。为了丰富手部信息,提出了基于 Laplace 逼近 Gaussian 过程的多分类算法,以实现基于手 部图像的指节识别。在类别信息无关联的假设基础上,将中层偏移测度特征的学习转化为对随 机量的学习;然后通过分析二值多分类高斯场上的后验计算,给出了基于 Laplace 逼近 Gaussian 过程的多分类高斯过程学习算法;通过构造中层随机信息的正定核函数,给出了基于 Laplace 的多分类高斯过程预测算法。最后,利用中层数据的分布学习与预测算法进行了指节图像训练 学习和固定阈值的图像识别。识别结果显示,该方法具有一定的指节识别能力。     

An electromyographic investigation of asymmetric lifting and moving of a load: II. Effects of load lift from knuckle to target shelves

The range of normalized electromyography (EMG) values when lifting and moving a 5.84 kg box was between 0.08 and 0.63. The right iliocostalis and the left erector spinae showed decreased muscle activity with increasing asymmetry. The right erector spinae and the left and the right external obliques showed significant increases in muscle activity at 90° asymmetry. The left and the right iliocostalis, the left erector spinae, and the right external oblique showed significant reductions in muscle activity for the sternum height when compared to the knuckle height. The right erector spinae exhibited a significant increase at reach height when compared to the knuckle height. The muscle activity for the knuckle lift was significantly lower than the floor lift. The average decrease in muscle activity was approximately 68%. Consequently, asymmetric lifting and moving tasks should be performed from a knuckle height to sternum height. © 2009 Wiley Periodicals, Inc.     

Ant colony optimization based fuzzy binary decision tree for bimodal hand knuckle verification system

In the recent trends of touch-less biometric authentication systems, hand knuckles from dorsal part of the hand is gaining popularity as a potential candidate for verification/recognition in variety of security applications. However, most of the available knuckle verification systems offer fixed security achieved for desired level of accuracy which cannot meet the varying levels of security requirements. This paper presents a bimodal knuckle verification system which is designed to meet a wide range of applications varying from civilian to high security regions. We use ant colony optimization (ACO) to choose the optimal fusion parameters corresponding to each level of security. The developed verification system utilizes fuzzy binary decision tree (FBDT) which is aimed at decision making in two classes: genuine (accept) and imposter (reject) using matching scores computed from the knuckle database. The FBDT is implemented using fuzzy Gini index for the selection of the tree nodes. The experiments are carried out on four publicly available HongKong PolyU knuckle databases named as: left index, right index, left middle and right middle with four bimodal systems: left–right index, left–right middle, left index–middle and right index–middle. The experimental results from these four bimodal knuckle databases validate the contributions of the proposed work.     

Validity of self-reported duration of work postures obtained by interview. MUSIC-Norrtälje Study Group

The aim of the study was to validate interview data concerning the duration of four work postures (1) sitting, (2) standing/walking with hands above shoulder level, (3) standing/walking with hands between shoulder and knuckle level, and (4) standing/walking with hands below knuckle level. The self-reported time spent in each posture was tested in relation to observations and technical measurements in 20 subjects during two full working days. The linear relationships between self-reports and observations were strong for the three postures; sitting (r2 = 0.55), hands above shoulder level (r2 = 0.58) and hands below knuckle level (r2 = 0.69). Thus, using this interview technique, self-reports concerning time spent in (1) sitting, (2) standing/walking with hands above shoulder level and, (3) standing/walking with hands below knuckle level may be accurate enough for studying these work postures in epidemiological studies.     

Validity of self-reported duration of work postures obtained by interview

The aim of the study was to validate interview data concerning the duration of four work postures (1) sitting, (2) standing/walking with hands above shoulder level, (3) standing/walking with hands between shoulder and knuckle level, and (4) standing/walking with hands below knuckle level. The self-reported time spent in each posture was tested in relation to observations and technical measurements in 20 subjects during two full working days. The linear relationships between self-reports and observations were strong for the three postures; sitting (r²=0.55), hands above shoulder level (r²=0.58) and hands below knuckle level (r²=0.69). Thus, using this interview technique, self-reports concerning time spent in (1) sitting, (2) standing/walking with hands above shoulder level and, (3) standing/walking with hands below knuckle level may be accurate enough for studying these work postures in epidemiological studies.     

Minimal acceptable handling time intervals for lifting and lowering tasks

A laboratory study was conducted to determine the minimal acceptable handling time interval (MAHTI) for 4 h of endurance manual materials handling tasks using a psychophysical approach. Nine experienced male subjects were recruited to perform manual materials handling tasks with three load weights (10, 15, and 20 kg) and six ranges (floor to knuckle, FK; knuckle to floor, KF; knuckle to shoulder, KS; shoulder to knuckle, SK; floor to shoulder, FS; shoulder to floor, SF). Subjects were asked to adjust the handling time interval to the minimum that they could perform without becoming strained, uncomfortable, tired, weakened, overheated, or out-of-breath for 4 h of the endurance handling period. The results showed that the heavier load produced a larger MAHTI, and the overall descending sequence of MAHTI values for the six ranges was FS, KS, SF, FK, KF, and SK. The MAHTI for each lifting task was significantly higher than that of its corresponding lowering task (p<0.05). All MAHTI data were further verified by tests run on an additional nine subjects; this showed that the psychophysically determined MAHTIs were appropriate with no apparent underestimation.     

人机交互装配中手部指形轮廓信息检测方法

人机交互装配系统中人体手部生物特征的识别与描述是手部装配动作描述与行为理解的基础。在肤色建模与识别、指节特征模型学习的基础上,通过肤色识别结果给出的模板轮廓坐标,计算轮廓链的局部转角比率和逐点向量积,描述基于逐点比较法的指节处的轮廓特征,结合手指生物结构的经验判据,对指尖与指根位置进行检测与推断,实现手指轮廓特征点的初步识别,通过指形拟合完成手指初步定位,为指形平面位姿分析提供基础。在手部多特征检测的基础上进行图像信息融合,结合指节学习模型与自适应阈值识别的指节位置检测,对指形的定位信息进行修正以提高检测的准确性。实验测试结果表明,该手部指形轮廓信息检测方法可行。     

This month Surprise and delight: statistical methods in ergonomics

Matching the job demands to a person's physical characteristics is an effective method of reducing the risk involved in a manual materials-handling task. Seventy-three male and 73 female industrial workers served as subjects. Maximum Acceptable Load of Lifting (MAL) was determined psychophysical under each combination of the following task conditions: six ranges of lift (floor to knuckle, floor to shoulder, floor to reach, knuckle to shoulder, knuckle to reach and shoulder to reach); four lifting frequencies (2,4,6 and 81/min); and three box sizes (30·48,45·72 and 60·96cm). A factor-score-based model (R² = 0-924) is developed in this paper based on 100 subjects for predicting an individual's MAL and for describing his/her physical characteristics in terms of strength and anthropometric scores. The model was validated on the remaining 46 subjects and was shown to be superior to the previously developed models from the same data set.     

Personal Authentication Using Finger Knuckle Surface

This paper investigates a new approach for personal authentication using fingerback surface imaging. The texture pattern produced by the finger knuckle bending is highly unique and makes the surface a distinctive biometric identifier. The finger geometry features can be simultaneously acquired from the same image at the same time and integrated to further improve the user-identification accuracy of such a system. The fingerback surface images from each user are normalized to minimize the scale, translation, and rotational variations in the knuckle images. This paper details the development of such an approach using peg-free imaging. The experimental results from the proposed approach are promising and confirm the usefulness of such an approach for personal authentication.      

Selection of optimized features for fusion of palm print and finger knuckle‐based person authentication

The impact of digital technology in biometrics is much more efficient at interpreting data than humans, which results in completely replacement of manual identification procedures in forensic science. Because the single modality‐based biometric frameworks limit performance in terms of accuracy and anti‐spoofing capabilities due to the presence of low quality data, therefore, information fusion of more than one biometric characteristic in pursuit of high recognition results can be beneficial. In this article, we present a multimodal biometric system based on information fusion of palm print and finger knuckle traits, which are least associated to any criminal investigation as evidence yet. The proposed multimodal biometric system might be useful to identify the suspects in case of physical beating or kidnapping and establish supportive scientific evidences, when no fingerprint or face information is present in photographs. The first step in our work is data preprocessing, in which region of interest of palm and finger knuckle images have been extracted. To minimize nonuniform illumination effects, we first normalize the detected circular palm or finger knuckle and then apply line ordinal pattern (LOP)‐based encoding scheme for texture enrichment. The nondecimated quaternion wavelet provides denser feature representation at multiple scales and orientations when extracted over proposed LOP encoding and increases the discrimination power of line and ridge features. To best of our knowledge, this first attempt is a combination of backtracking search algorithm and 2D²LDA has been employed to select the dominant palm and knuckle features for classification. The classifiers output for two modalities are combined at unsupervised rank level fusion rule through Borda count method, which shows an increase in performance in terms of recognition and verification, that is, 100% (correct recognition rate), 0.26% (equal error rate), 3.52 (discriminative index), and 1,262 m (speed).     

The psychophysical lifting capacities of Chinese subjects

The psychophysical lifting capacity (MAWL) of twelve subjects was determined in this study. The subjects were all young Chinese males who performed lifting tasks in three lifting ranges (floor to knuckle, floor to shoulder, and knuckle to shoulder) and four lifting frequencies (one-time maximum, 1 lift/min, 4 lifts/min, and 6 lifts/min). The oxygen uptake (1/min) and heart rate (beats/min) were recorded while subjects were lifting. Upon completion of each lifting task, the subjects were required to rate their perceived exertion levels. The statistical analyses results indicated the following. Chinese subjects have smaller body size and MAWLs compared with past studies using the US population. The MAWLs decreased with an increase in lifting frequencies. The decrements of MAWL due to lifting frequencies were in agreement with the results of past studies. However, there were larger decreases due to lifting ranges. The MAWLs of the floor to knuckle height lift were the largest, followed by the MAWLs of the floor to shoulder height lift, and the MAWLs of the knuckle to shoulder height lift. The measured physiological responses were considered similar to those obtained in past studies. Subjects' perceived stress levels increased with the lifting frequency and the upper extremities received the most stress for the total range of lifting tasks. The comparisons of the Chinese MAWLs with the NIOSH lifting guidelines for limits (AL and MPL) indicated that the vertical discounting factor in the guidelines should be modified before the NIOSH limits can be applied to non-Western populations.     

Finger knuckle biometric feature selection based on the FIS_DE optimization algorithm

Neural Computing and Applications - In recent years, the hand-based biometric system has received significant attention in identifying a person. In the hand-based biometric, the finger knuckle...     

Psychophysical modelling for combined manual materials-handling activities

Most psychophysical studies in manual materials handling (MMH)are involved only with single MMH activities, i.e. lifting, lowering, carrying, holding, pushing or pulling. Very little research has been reported on the determination of operator capacities for combinations of MMH activities (e.g. lifting a box, then carrying the box, or carrying a box, then lowering the box). These kinds of combined activities are prevalent in industry and in our daily lives. The objective of this study was to utilize the psychophysical approach to examine the effects of combinations of lifting, carrying and lowering activities. Twelve male students served as subjects for the study. The capacities that were determined as the maximum acceptable workloads for a 1-h work period for four individual MMH activities—lifting from floor to knuckle height (LFK), lifting from knuckle to shoulder height (LKS), lowering from knuckle to floor height (LOW) and carrying for 3·4 m (C) —and three combined MMH activities—LFK + C, LFK + C + LKS and LFK + C + LOW—were determined psychophysically under three frequency conditions: one time maximum, one handling per minute and six handlings per minute. Combined MMH capacities models were developed using the following three methods: a limiting individual MMH capacity, isoinertial 1·83-m maximum strength and fuzzy-set theory. The advantages and disadvantages of different models were discussed.     

基于指部关联特征的多模态图像采集系统

针对基于单个生物特征的身份认证安全性和稳定性不足的问题,设计了基于指部关联特征的多模态图像采集系统,采用单个双波段摄像头分时采集同一根手指的指纹、指节纹和指静脉图像。指纹和指节纹采用非接触反射采集方式,指静脉采用单侧近红外光源与反射镜面相结合的透射采集方式,并根据静脉图像质量评价动态调控光源,根据特征点信息量动态调整各个特征的权重。实验结果表明,该多模态采集系统在认证通过率、误识率和拒登率等指标都优于指纹或指静脉的单模态采集系统,认证通过率达到99.1%,误识率为0.000 1%,不存在拒登现象。     

Psychophysically determined infrequent lifting capacity of Chinese participants.

The aim of this study was to investigate the psychophysical infrequent lifting capacity (maximum acceptable weight of lift, MAWL) for a Chinese population. A nested-factorial experimental design with a participant factor nested within gender was employed. Forty-one Chinese participants (29 males, 12 females) participated in the study. Two frequencies (one lift every 8 h and one lift every 5 min) and six lifting heights (floor to knuckle, floor to shoulder, floor to reach, knuckle to shoulder, knuckle to reach, shoulder to reach) were evaluated. The results are compared with prior studies and they lead to the following conclusions. (1) The MAWLs were significantly affected by both the lifting frequency and lifting height. For lifting frequency, the MAWLs decreased markedly by nearly 30% from one lift every 8 h to one lift every 5 min. For lifting height, the MAWLs of the F-K was the greatest of all six lifting heights, followed by F-S, K-S, F-R, K-R and S-R was the smallest. (2) The MAWLs for Chinese females are significantly lower, but proportionately similar, to the MAWLs for Chinese males. (3) The Chinese participants had smaller capacities compared with the Occidental participants, and the rate of decrease in MAWL for the Chinese participants was much sharper than that of Occidental participants. (4) The MAWL of Chinese females was about 54-58% of the Chinese males, it is somewhat lower than those of 60-70% of the Occidental participants.     

Relationship between posture and lifting capacity

This study investigates the effect of changes in posture caused by wearing high-heeled shoes on the maximum lifting capacity. Nine female college students, ages 20 to 25 years, participated in this study. Three heel heights (flat, 5 cm and 7.6 cm), two lifting heights (floor to knuckle and knuckle to shoulder), and lifting frequency of 4 per minute were examined. The results indicate that a significant difference exists between MAWOL with flats and that with 7.6 cm heels for both lifting heights. Subjects lifted 21.5% less weight using 7.6 cm heels than wearing flats. No significant difference was found between MAWOL with flats and 5 cm high heels. In addition, in evaluating the tasks subjectively, the subjects reported that they experienced a stress ontheir legs when lifting with 5 cm and 7.6 cm high-heeled shoes. The conclusion of this study indicates that a change in posture affects lifting capacity, and individuals should adjust their predetermined MAWOL while wearing high-heeled shoes.     

汽车转向节差压铸造过程温度场精确建模和计算

为确保差压铸造汽车转向节的性能满足要求,基于实际铸造工艺对铝合金AlSi7Mg0.3的热物理性能进行精确测量和修正,建立仿真度较高的转向节差压铸造过程物理模型,计算系统不同部位的温度演化过程以及电偶测量值与实际温度间的差异。校准后的电偶实测温度曲线与计算温度曲线对比表明：计算所得的温度曲线与实测曲线形状完全一致,波峰、波谷处数值模拟的准确率达到93%以上。在此基础上进行铸件缩孔疏松预测,模拟结果与实际解剖结果一致,可进一步提高差压铸造温度场计算和测量的精确度,提高铸造缺陷预测的准确性。     

Psychophysical lifting capabilities for overreach heights

Overreach height, in this study, is defined as the maximum reach height of individuals measured to the top of the cut-out box-handles while subjects stand with their heels raised. Since such postures are inherently unstable, knowing how much weight individuals are willing to lift across overreach lifting heights is important. Ten young adult male students (mean age 25·9 years, mean weight 70·8 kg and mean height 175 cm) voluntarily participated in a study designed to investigate the effect of lifting heights above reach height on the maximum acceptable weights of lift. The weight was lifted using a ‘free-style’ technique in the mid-sagittal plane from the floor, knuckle and shoulder heights to overreach heights (individuals stand with their heels raised to deposit the load). The maximum acceptable weight of lift, on the average, declined by approximately 14%, compared with the maximum acceptable weight of lift for reach heights, when the box was lifted to overreach heights. The magnitude of decline in the maximum acceptable weight was highest for the floor to overreach height compared with the knuckle to overreach and shoulder to overreach lifting heights.