Prediction accuracy in estimating joint angle trajectories using a video posture coding method for sagittal lifting tasks

This study investigated prediction accuracy of a video posture coding method for lifting joint trajectory estimation. From three filming angles, the coder selected four key snapshots, identified joint angles and then a prediction program estimated the joint trajectories over the course of a lift. Results revealed a limited range of differences of joint angles (elbow, shoulder, hip, knee, ankle) between the manual coding method and the electromagnetic motion tracking system approach. Lifting range significantly affected estimate accuracy for all joints and camcorder filming angle had a significant effect on all joints but the hip. Joint trajectory predictions were more accurate for knuckle-to-shoulder lifts than for floor-to-shoulder or floor-to-knuckle lifts with average root mean square errors (RMSE) of 8.65°, 11.15° and 11.93°, respectively. Accuracy was also greater for the filming angles orthogonal to the participant's sagittal plane (RMSE = 9.97°) as compared to filming angles of 45° (RMSE = 11.01°) or 135° (10.71°). The effects of lifting speed and loading conditions were minimal. To further increase prediction accuracy, improved prediction algorithms and/or better posture matching methods should be investigated.

Statement of Relevance: Observation and classification of postures are common steps in risk assessment of manual materials handling tasks. The ability to accurately predict lifting patterns through video coding can provide ergonomists with greater resolution in characterising or assessing the lifting tasks than evaluation based solely on sampling with a single lifting posture event.     

Semantic image segmentation using low-level features and contextual cues

Semantic image segmentation aims to partition an image into non-overlapping regions and assign a pre-defined object class label to each region. In this paper, a semantic method combining low-level features and high-level contextual cues is proposed to segment natural scene images. The proposed method first takes the gist representation of an image as its global feature. The image is then over-segmented into many super-pixels and histogram representations of these super-pixels are used as local features. In addition, co-occurrence and spatial layout relations among object classes are exploited as contextual cues. Finally the features and cues are integrated into the inference framework based on conditional random field by defining specific potential terms and introducing weighting functions. The proposed method has been compared with state-of-the-art methods on the MSRC database, and the experimental results show its effectiveness.     

Dimensionality reduction of image features using the canonical contextual correlation projection

A linear, discriminative, supervised technique for reducing feature vectors extracted from image data to a lower-dimensional representation is proposed. It is derived from classical linear discriminant analysis (LDA), extending this technique to cases where there is dependency between the output variables, i.e., the class labels, and not only between the input variables. (The latter can readily be dealt with in standard LDA.) The novel method is useful, for example, in supervised segmentation tasks in which high-dimensional feature vectors describe the local structure of the image.

The principal idea is that where standard LDA merely takes into account a single class label for every feature vector, the new technique incorporates class labels of its neighborhood in the analysis as well. In this way, the spatial class label configuration in the vicinity of every feature vector is accounted for, resulting in a technique suitable for, e.g. image data.

This extended LDA, that takes spatial label context into account, is derived from a formulation of standard LDA in terms of canonical correlation analysis. The novel technique is called the canonical contextual correlation projection (CCCP).

An additional drawback of LDA is that it cannot extract more features than the number of classes minus one. In the two-class case this means that only a reduction to one dimension is possible. Our contextual LDA approach can avoid such extreme deterioration of the classification space and retain more than one dimension.

The technique is exemplified on a pixel-based medical image segmentation problem in which it is shown that it may give significant improvement in segmentation accuracy.     

Development of a system usability assessment procedure using oculo-motors for input operation

This paper investigates the relationship between oculo-motors, such as eye-movement and pupillary change, and the conventional subjective “usability” index, using time-domain and frequency-domain approaches, with the objective to determine the possibility of evaluating interaction through oculo-motors. An evaluation experiment was conducted by operating a target on a computer display using mouse, keyboard and key pad as input devices. The results show that there is a significant correlation between pupil size and SU-score, which is an established subjective evaluation index for system usability. Cross spectrum densities (CSD) between horizontal and vertical eye-movements and coherence as standardized CSD also significantly correlate with the results of the SU-scores and error rates. To determine the frequency range of CSD and coherence for usability assessment, frequency components used as factors were extracted using factor analysis. According to the correlation coefficients between these and the performance of factor scores for predicting the conventional metrics, factor scores of CSD are better indices for assessing usability than factor scores of coherence. These two results suggest that pupil size and index of eye-movement as oculo-motor indices based on time-domain and frequency-domain approaches can provide information about a system’s overall usability regarding the input operation task.     

Improving the classification accuracy of streaming data using SAX similarity features

The classification accuracy of time series is highly dependent on the quality of used features. In this study, features of new type, called SAX (Symbolic Aggregate approXimation) similarity features, are presented. SAX similarity features are a combination of the traditional statistical number-based and the template-based classification. SAX similarity features are obtained from the data of the time window by first transforming the time series into a discrete presentation using SAX. Then the similarity between this SAX presentation and predefined SAX templates are calculated, and these similarity values are considered as SAX similarity features. The functioning of these features was tested using five different activity data sets collected using wearable inertial sensors and five different classifiers. The results show that the recognition rates calculated using SAX similarity features together with traditional features are much better than those obtained employing traditional features only. In 20 tested cases out of 23, the improvement is statistically significant according to the paired t-test.     

The effect of banner animation on fixation behavior and recall performance in search tasks

Previous findings suggested that banner ads have little or no impact on perceptual behavior and memory performance in search tasks, but only in browsing paradigms. This assumption is not supported by the present eye-tracking study. It investigates whether task-related selective attention is disrupted depending on the animation intensity of banner ads when users are in a search mode as well as the impact of banner animation on perceptual and memory performance.     

GCP acquisition using simulated SAR and evaluation of GCP matching accuracy with texture features

Abstract

A method for ground control point (GCP) acquisition using a simulated synthetic aperture radar (SAR) image derived from a digital elevation model (DEM) is proposed. Also proposed is a method for the evaluation of the accuracy of GCP matching with texture features from a reference GCP chip image. Results from experiments with simulated GCP chip images as reference images and geometrically distorted GCP chip images, derived using simulated SAR images as current images, show good coincidence with GCP matching accuracy in terms of pixel distances between matched GCP points in reference and current chip images and texture features. In particular, the correlation coefficient between the angular second moment (ASM) and the matching accuracy is 0.757, followed by Ent (entropy), Horn (homoginiety), Dis (dissimilarity), Con (contrast), Var (variability coefficient) and Chi (Chi-square), for the skewed images while for the rotated images, Dis shows the highest correlation of 0.628, followed by Horn, Asm, Con, Ent, Chi and Var. Based on the proposed methods, one can easily generate GCP chips images from a DEM, and also evaluate GCP matching accuracy with texture features of simulated SAR from a GCP chip image.     

Speed and accuracy effects of fingers and dexterity in 5-choice reaction tasks

In two experiments we studied the influence of dexterity (controls, typists, pianists) and of differences between the five fingers of the dominant hand on speed and accuracy in a 5-choice reaction task. We used five coloured squares (Blue, Green, Yellow, Red, White) as stimuli, randomly varying foreperiods (3-10 s) and an intermediate stimulus-response-compatibility. The results reported here were independent of sex, foreperiods, and colour of stimuli.

In Experiment 1 (JV=168) with three groups (controls, typists, pianists) thumb and little finger showed significantly shorter reaction times than did index, middle, and ring finger. This difference did not interact with dexterity. Averaged across all the fingers typists and controls did not differ. Pianists showed significantly shorter reaction times than these two groups. Experiment 2 (N = 40) replicated all these findings

In both experiments groups did not differ in speed-accuracy-trade-off. The accuracy of the fingers was independent of dexterity. In all three groups the little finger showed the lowest rate of false alarms and the highest degree of reliability. The frequency distribution of finger confusions in all groups was in accordance with the spatial proximity of fingers: the closer their proximity the higher were the rates of confusion. In addition, these distributions showed an asymmetry across all three groups. The finger next to the 'correct’ finger in direction towards the thumb showed the highest false alarm rate in each case.     

Speed and accuracy effects of fingers and dexterity in 5-choice reaction tasks

In two experiments we studied the influence of dexterity (controls, typists, pianists) and of differences between the five fingers of the dominant hand on speed and accuracy in a 5-choice reaction task. We used five coloured squares (Blue, Green, Yellow, Red, White) as stimuli, randomly varying foreperiods (3-10s) and an intermediate stimulus-response-compatibility. The results reported here were independent of sex, foreperiods, and colour of stimuli. In Experiment 1 (N = 168) with three groups (controls, typists, pianists) thumb and little finger showed significantly shorter reaction times than did index, middle, and ring finger. This difference did not interact with dexterity. Averaged across all the fingers typists and controls did not differ. Pianists showed significantly shorter reaction times than these two groups. Experiment 2 (N = 40) replicated all these findings. In both experiments groups did not differ in speed-accuracy-trade-off. The accuracy of the fingers was independent of dexterity. In all three groups the little finger showed the lowest rate of false alarms and the highest degree of reliability. The frequency distribution of finger confusions in all groups was in accordance with the spatial proximity of fingers: the closer their proximity the higher were the rates of confusion. In addition, these distributions showed an asymmetry across all three groups. The finger next to the 'correct' finger in direction towards the thumb showed the highest false alarm rate in each case.     

Self-learning classification of radar features for scene understanding

Autonomous driving is a challenging problem in mobile robotics, particularly when the domain is unstructured, as in an outdoor setting. In addition, field scenarios are often characterized by low visibility as well, due to changes in lighting conditions, weather phenomena including fog, rain, snow and hail, or the presence of dust clouds and smoke. Thus, advanced perception systems are primarily required for an off-road robot to sense and understand its environment recognizing artificial and natural structures, topology, vegetation and paths, while ensuring, at the same time, robustness under compromised visibility. In this paper the use of millimeter-wave radar is proposed as a possible solution for all-weather off-road perception. A self-learning approach is developed to train a classifier for radar image interpretation and autonomous navigation. The proposed classifier features two main stages: an adaptive training stage and a classification stage. During the training stage, the system automatically learns to associate the appearance of radar data with class labels. Then, it makes predictions based on past observations. The training set is continuously updated online using the latest radar readings, thus making it feasible to use the system for long range and long duration navigation, over changing environments. Experimental results, obtained with an unmanned ground vehicle operating in a rural environment, are presented to validate this approach. A quantitative comparison with laser data is also included showing good range accuracy and mapping ability as well. Finally, conclusions are drawn on the utility of millimeter-wave radar as a robotic sensor for persistent and accurate perception in natural scenarios.     

Prediction of track purity and track accuracy in dense targetenvironments

This note is concerned with performance prediction of multiple-target tracking systems, particularly in data-rich, high-target-density environments. Two simple but useful expressions are described: (i) the exponential law for predicting average probability of correct data association, and (ii) the linear-times-exponential law for calculating the effective measurement error variances with possible data misassociation. Estimation of two key multiple-target tracking performance measures, i.e., track purity and track accuracy, are then derived using these two laws     

A comparison of contextual classification methods using Landsat TM

The performance of contextual classification methods is evaluated using Landsat TM data. Classes of pixels adjacent to the pixel to be classified are assumed to be conditionally independent given the class of the pixel to be classified. An assumption of autocorrelated spectral reflectance is made in three of the methods. Methods that utilize information from one image and images from two different occasions are compared.Our results indicate that an autocorrelation method utilizing images from two different occasions performs optimally.     

Prediction of epileptic seizures from spectral features of intracranial eeg recordings using deep learning approach

Epilepsy is a prevalent neurological disorder, which disturbs the lives of millions of people worldwide owing to the onset of abrupt seizures. The forecasting of seizures could help in protecting their lives by alerts or in clinical operations during epilepsy surgeries. The present paper addresses this problem by proposing a deep learning framework for prediction of epileptic seizures using intracranial EEG (iEEG) recordings. This framework performs filtering and segmentation of iEEG signals into 10s, 20s, 30s, 40s, 50s and 60s duration segments. These segments are further resolved into eight distinct spectral bands corresponding to delta, theta, alpha, beta and gamma sub-bands with frequency-domain transformation. Then, mean amplitude and band power features are extracted from each band, which are provided to convolutional neural network (CNN) and long short-term memory network (LSTM) algorithms for classification. The simulation results of the proposed CNN model exhibit higher performance with average accuracy, sensitivity, specificity, AUC and F1 score of 94.74%, 95.8%, 94.46%, 95.13% and 94.75% respectively for iEEG segments of 40s duration. Thus, the performance analysis and comparison with existing literature unveil that the proposed CNN model is an optimal approach for accurate and real-time prediction of epileptic seizures.

     

Classification of mental tasks using stockwell transform

In recent years, various physiological signal based rehabilitation systems have been developed for the physically disabled in which electroencephalographic (EEG) signal is one among them. The efficiency of such a system depends upon the signal processing and classification algorithms. In order to develop an EEG based rehabilitation or assistive system, it is necessary to develop an effective EEG signal processing algorithm. This paper proposes Stockwell transform (ST) based analysis of EEG dynamics during different mental tasks. EEG signals from Keirn and Aunon database were used in this study. Three classifiers were employed such as k-means nearest neighborhood (kNN), linear discriminant analysis (LDA) and support vector machine (SVM) to test the strength of the proposed features. Ten-fold cross validation method was used to demonstrate the consistency of the classification results. Using the proposed method, an average accuracy ranging between 84.72% and 98.95% was achieved for multi-class problems (five mental tasks).     

Ergonomic evaluation of hospital bed design features during patient handling tasks

Patient handling tasks (e.g., transportation and repositioning) are important causes of musculoskeletal disorders among healthcare workers. The purpose of this study was to evaluate, during two patient handling tasks, the physical demands resulting from alternative hospital bed design features. Twenty-four novice participants were involved in two laboratory-based studies. The effects of a steering lock and adjustable push height were evaluated during a patient transportation task using perceptual responses and measures of performance and physical demands, and the effect of a bed contour feature was determined based on patient sliding distance during repeated bed raising/lowering. Use of the steering lock reduced the number of adjustments during bed maneuvering by 28% and decreased ratings of physical demands. Use of the adjustable push height reduced shoulder moments by 30%. With the contour feature, patient sliding distance was reduced by ∼40% over 12 raise/lower cycles. These results suggest that the steering lock and adjustable push height features can reduce physical demands placed on healthcare workers during patient transportation tasks. Although patient sliding distance was reduced using the contour feature, assessing direct effects of this feature on physical demands (e.g., reduced need for workers to reposition patients) will require further investigation.

Relevance to industry

Hospital bed design features have the potential to reduce physical demands required of healthcare workers, yet there have been only limited empirical studies of these. Findings of the two current studies suggest that proactive ergonomic considerations in hospital bed design can reduce these physical demands.     

View independent face detection based on horizontal rectangular features and accuracy improvement using combination kernel of various sizes

This paper proposes a view independent face detection method based on horizontal rectangular features, and accuracy improvement by combining kernels of various sizes. Since the view changes of faces induce large variation in appearance in the horizontal direction, local kernels are applied to horizontal rectangular regions to model such appearance changes. Local kernels are integrated by summation, and then used as a summation kernel for support vector machine (SVM). View independence is shown to be realized by the integration of local horizontal rectangular kernels. However, in general, local kernels (features) of various sizes have different similarity measures, such as detailed and rough similarity, and thus their error patterns are different. If the local and global kernels are combined well, the generalization ability is improved. This research demonstrates the comparative effectiveness of combining the global kernel and local kernels of various sizes as a summation kernel for SVM against use of only the global kernel, only the combination of local kernels and Adaboost with SVMs with a kind of local kernel.