Search results

In this paper, an emerging state-of-the-art machine intelligence technique called the Hierarchical Temporal Memory (HTM) is applied to the task of short-term load forecasting (STLF). A HTM Spatial Pooler (HTM-SP) stage is used to continually form sparse distributed representations (SDRs) from a univariate load time series data, a temporal aggregator is used to transform the SDRs into a sequential bivariate representation space and an overlap classifier makes temporal classifications from the bivariate SDRs through time. The comparative performance of HTM on several daily electrical load time series data including the Eunite competition dataset and the Polish power system dataset from 2002 to 2004 are presented. The robustness performance of HTM is also further validated using hourly load data from three more recent electricity markets. The results obtained from experimenting with the Eunite and Polish dataset indicated that HTM will perform better than the existing techniques reported in the literature. In general, the robustness test also shows that the error distribution performance of the proposed HTM technique is positively skewed for most of the years considered and with kurtosis values mostly lower than a base value of 3 indicating a reasonable level of outlier rejections.

Geographical conditions monitoring (GCM) has elicited significant concerns from the Chinese Government and is closely related to the “Digital China” program. This research aims to focus on object-based change detection (OBCD) methods integrating very-high-resolution (VHR) imagery and vector data for GCM.

The main content of this paper is as follows: a multi-resolution segmentation (MRS) algorithm is proposed for obtaining homogeneous and contiguous image objects in two phases; a post-classification comparison (PCC) method based on the nearest neighbor algorithm and an image-object analysis (IOA) technique based on a differential entropy algorithm are used to improve the accuracy of the change detection; and a vector object-based accuracy assessment method is proposed.

Results show that image objects obtained using the MRS algorithm attain the objectives of the “same spectrum within classes” and “different spectrum among classes”. Moreover, the two OBCD methods can detect over 85 per cent of the changed regions. The PCC strategy can obtain the categories of image objects with a high degree of precision. The IOA technique is easy to use and largely automated.

On the basis of the VHR satellite imagery and vector data, the above methods can effectively and accurately provide technical support for GCM implementation.

Road accidents, an inadvertent mishap can be detected automatically and alerts sent instantly with the collaboration of image processing techniques and on-road video surveillance systems. However, to rely exclusively on visual information especially under adverse conditions like night times, dark areas and unfavourable weather conditions such as snowfall, rain, and fog which result in faint visibility lead to incertitude. The main goal of the proposed work is certainty of accident occurrence.

The authors of this work propose a method for detecting road accidents by analyzing audio signals to identify hazardous situations such as tire skidding and car crashes. The motive of this project is to build a simple and complete audio event detection system using signal feature extraction methods to improve its detection accuracy. The experimental analysis is carried out on a publicly available real time data-set consisting of audio samples like car crashes and tire skidding. The Temporal features of the recorded audio signal like Energy Volume Zero Crossing Rate 28ZCR2529 and the Spectral features like Spectral Centroid Spectral Spread Spectral Roll of factor Spectral Flux the Psychoacoustic features Energy Sub Bands ratio and Gammatonegram are computed. The extracted features are pre-processed and trained and tested using Support Vector Machine (SVM) and K-nearest neighborhood (KNN) classification algorithms for exact prediction of the accident occurrence for various SNR ranges. The combination of Gammatonegram with Temporal and Spectral features of the validates to be superior compared to the existing detection techniques.

Temporal, Spectral, Psychoacoustic features, gammetonegram of the recorded audio signal are extracted. A High level vector is generated based on centroid and the extracted features are classified with the help of machine learning algorithms like SVM, KNN and DT. The audio samples collected have varied SNR ranges and the accuracy of the classification algorithms is thoroughly tested.

Denoising of the audio samples for perfect feature extraction was a tedious chore.

The existing literature cites extraction of Temporal and Spectral features and then the application of classification algorithms. For perfect classification, the authors have chosen to construct a high level vector from all the four extracted Temporal, Spectral, Psycho acoustic and Gammetonegram features. The classification algorithms are employed on samples collected at varied SNR ranges.

Accurate segmentation of artificial assembly action is the basis of autonomous industrial assembly robots. This paper aims to study the precise segmentation method of manual assembly action.

In this paper, a temporal-spatial-contact features segmentation system (TSCFSS) for manual assembly actions recognition and segmentation is proposed. The system consists of three stages: spatial features extraction, contact force features extraction and action segmentation in the temporal dimension. In the spatial features extraction stage, a vectors assembly graph (VAG) is proposed to precisely describe the motion state of the objects and relative position between objects in an RGB-D video frame. Then graph networks are used to extract the spatial features from the VAG. In the contact features extraction stage, a sliding window is used to cut contact force features between hands and tools/parts corresponding to the video frame. Finally, in the action segmentation stage, the spatial and contact features are concatenated as the input of temporal convolution networks for action recognition and segmentation. The experiments have been conducted on a new manual assembly data set containing RGB-D video and contact force.

In the experiments, the TSCFSS is used to recognize 11 kinds of assembly actions in demonstrations and outperforms the other comparative action identification methods.

A novel manual assembly actions precisely segmentation system, which fuses temporal features, spatial features and contact force features, has been proposed. The VAG, a symbolic knowledge representation for describing assembly scene state, is proposed, making action segmentation more convenient. A data set with RGB-D video and contact force is specifically tailored for researching manual assembly actions.

The purpose of this paper is to propose a gait analysis technique that aims to identify differences and similarities in gait performance between three different assistive devices (ADs).

Two feature reduction techniques, linear principal component analysis (PCA) and nonlinear kernel-PCA (KPCA), are expanded to provide a comparison of the spatio-temporal, symmetrical indexes and postural control parameters among the three different ADs. Then, a multiclass support vector machine (MSVM) with different approaches is designed to evaluate the potential of PCA and KPCA to extract relevant gait features that can differentiate between ADs.

Results demonstrated that symmetrical indexes and postural control parameters are better suited to provide useful information about the different gait patterns that total knee arthroplasty (TKA) patients present when walking with different ADs. The combination of KPCA and MSVM with discriminant functions (MSVM DF) resulted in a noticeably improved performance. Such combination demonstrated that, with symmetric indexes and postural control parameters, it is possible to extract with high-accuracy nonlinear gait features for automatic classification of gait patterns with ADs.

The information obtained with the proposed technique could be used to identify benefits and limitations of ADs on the rehabilitation process and to evaluate the benefit of their use in TKA patients.

The purpose of this paper is to identify, classify and analyse temporality in information systems development (ISD) literature.

The authors address the temporality and ISD research gap by using a framework – which classifies time into three categories: conceptions of time, mapping activities to time and actors relating to time. The authors conduct a systematic literature review which investigates time in ISD within the Senior Scholars' Basket, Information Technology & People (IT&P), and top two information systems conferences over the past 20 years. The search strategy resulted in 9,850 studies of which 47 were identified as primary papers.

The results reveal that ISD research is ill equipped for contemporary thinking around time. This systematic literature review (SLR) contributes to ISD by finding the following gaps in the literature: (1) clock time is dominant and all other types of time are under-researched; (2) contributions to mapping activities to time is lacking and existing studies focus on single ISD projects rather multiple complex ISD projects; (3) research on actors relating to time is lacking; (4) existing ISD studies which contribute to temporal characteristics are fragmented and lack integration with other categories of time and (5) ISD methodology papers lack contributions to temporal characteristics and fail to acknowledge and contribute to time as a multifaceted interrelated concept.

This work has developed the first SLR on temporality in ISD. This study provides a starting point for ISD researchers and ISD practitioners to test commonly held temporal assumptions of ISD researchers and practitioners.

This study aims to reduce data bias during human activity and increase the accuracy of activity recognition.

A convolutional neural network and a bidirectional long short-term memory model are used to automatically capture feature information of time series from raw sensor data and use a self-attention mechanism to learn select potential relationships of essential time points. The proposed model has been evaluated on six publicly available data sets and verified that the performance is significantly improved by combining the self-attentive mechanism with deep convolutional networks and recursive layers.

The proposed method significantly improves accuracy over the state-of-the-art method between different data sets, demonstrating the superiority of the proposed method in intelligent sensor systems.

Using deep learning frameworks, especially activity recognition using self-attention mechanisms, greatly improves recognition accuracy.

Location‐prediction enables the next generation of location‐based applications. The purpose of this paper is to provide a historical summary of research in personal location‐prediction. Location‐prediction began as a tool for network management, predicting the load on particular cellular towers or WiFi access points. With the increasing popularity of mobile devices, location‐prediction turned personal, predicting individuals' next locations given their current locations.

This paper includes an overview of prediction techniques and reviews several location‐prediction projects comparing the raw location data, feature extraction, choice of prediction algorithms and their results.

A new trend has emerged, that of employing additional context to improve or expand predictions. Incorporating temporal information enables location‐predictions farther out into the future. Appending place types or place names can improve predictions or develop prediction applications that could be used in any locale. Finally, the authors explore research into diverse types of context, such as people's personal contacts or health activities.

This overview provides a broad background for future research in prediction.

The purpose of this paper is to develop a new method to extract medical temporal information from online health communities.

The authors trained a conditional random-filed model for the extraction of temporal expressions. The temporal relation identification is considered as a classification task and several support vector machine classifiers are built in the proposed method. For the model training, the authors extracted some high-level semantic features including co-reference relationship of medical concepts and the semantic similarity among words.

For the extraction of TIMEX, the authors find that well-formatted expressions are easy to recognize, and the main challenge is the relative TIMEX such as “three days after onset”. It also shows the same difficulty for normalization of absolute date or well-formatted duration, whereas frequency is easier to be normalized. For the identification of DocTimeRel, the result is fairly well, and the relation is difficult to identify when it involves a relative TIMEX or a hypothetical concept.

The authors proposed a new method to extract temporal information from the online clinical data and evaluated the usefulness of different level of syntactic features in this task.

The purpose of this paper to classify a set of Turkish sign language (TSL) gestures by posture labeling based finite-state automata (FSA) that utilize depth values in location-based features. Gesture classification/recognition is crucial not only in communicating visually impaired people but also for educational purposes. The paper also demonstrates the practical use of the techniques for TSL.

Gesture classification is based on the sequence of posture labels that are assigned by location-based features, which are invariant under rotation and scale. Grid-based signing space clustering scheme is proposed to guide the feature extraction step. Gestures are then recognized by FSA that process temporally ordered posture labels.

Gesture classification accuracies and posture labeling performance are compared to k-nearest neighbor to show that the technique provides a reasonable framework for recognition of TSL gestures. A challenging set of gestures is tested, however the technique is extendible, and extending the training set will increase the performance.

The outcomes can be utilized as a system for educational purposes especially for visually impaired children. Besides, a communication system would be designed based on this framework.

The posture labeling scheme, which is inspired from keyframe labeling concept of video processing, is the original part of the proposed gesture classification framework. The search space is reduced to single dimension instead of 3D signing space, which also facilitates design of recognition schemes. Grid-based clustering scheme and location-based features are also new and depth values are received from Kinect. The paper is of interest for researchers in pattern recognition and computer vision.