In recent years, location-based technologies for ubiquitous environments have aimed to realize services tailored to each purpose based on information about an individual's current location. To establish such advanced location-based services, an estimation technology that can accurately recognize and predict the movements of people and objects is necessary. Although global positioning system (GPS) has already been used as a standard for outdoor positioning technology and many services have been realized, several techniques using conventional wireless sensors such as Wi-Fi, RFID, and Bluetooth have been considered for indoor positioning technology. However, conventional wireless indoor positioning is prone to the effects of noise, and the large range of estimated indoor locations makes it difficult to identify human activities precisely. We propose a method to mine user activity patterns from time-series data of user's locationss in an indoor environment using ultra-wideband (UWB) sensors. An UWB sensor is useful for indoor positioning due to its high noise immunity and measurement accuracy, however, to our knowledge, estimation and prediction of human indoor activities using UWB sensors have not yet been addressed. The proposed method consists of three steps: 1) obtaining time-series data of the user's location using a UWB sensor attached to the user, and then estimating the areas where the user has stayed; 2) associating each area of the user's stay with a nearby landmark of activity and assigning indoor activities; and 3) mining the user's activity patterns based on the user's indoor activities and their transitions. We conducted experiments to evaluate the proposed method by investigating the accuracy of estimating the user's area of stay using a UWB sensor and observing the results of activity pattern mining applied to actual laboratory members over 30-days. The results showed that the proposed method is superior to a comparison method, Time-based clustering algorithm, in estimating the stay areas precisely, and that it is possible to reveal the user's activity patterns appropriately in the actual environment.

In recent years, with the aging of society, many kinds of research have been actively conducted to recognize human activity in a home to watch over the elderly. Multiple sensors for activity recognition are used. However, we need to consider privacy when using these sensors. One of the candidates of the sensors that keep privacy is a sound sensor. MFCC (Mel-Frequency Cepstral Coefficient) is widely used as a feature extraction algorithm for voice recognition. However, it is not suitable to apply conventional MFCC to activity recognition by sounds of daily life. We denote “sounds of daily life” as “life sounds” simply in this paper. The reason is that conventional MFCC does not extract well several features of life sounds that appear at high frequencies. This paper proposes the improved MFCC and reports the evaluation results of activity recognition by machine learning SVM (Support Vector Machine) using features extracted by improved MFCC.

The triaxial accelerometer is one of the most important sensors for human activity recognition (HAR). It has been observed that the relations between the axes of a triaxial accelerometer plays a significant role in improving the accuracy of activity recognition. However, the existing research rarely focuses on these relations, but rather on the fusion of multiple sensors. In this paper, we propose a data fusion-based convolutional neural network (CNN) approach to effectively use the relations between the axes. We design a single-channel data fusion method and multichannel data fusion method in consideration of the diversified formats of sensor data. After obtaining the fused data, a CNN is used to extract the features and perform classification. The experiments show that the proposed approach has an advantage over the CNN in accuracy. Moreover, the single-channel model achieves an accuracy of 98.83% with the WISDM dataset, which is higher than that of state-of-the-art methods.

This study designs a low-cost portable device that functions as a coaching assistant system which can support table tennis practice. Although deep learning technology is a promising solution to realizing human activity recognition, we propose using cosine similarity in making inferences. Our experiments show that the cosine similarity based inference can be a good alternative to the deep learning based inference for the assistant system when resources are limited.

We propose a class-incremental learning framework for human activity recognition based on the Bag-of-Sequencelets model (BoS). The framework updates learned models efficiently without having to relearn them when training data of new classes are added. In this framework, all types of features including hand-crafted features and Convolutional Neural Networks (CNNs) based features and combinations of those features can be used as features for videos. Compared with the original BoS, the new framework can reduce the learning time greatly with little loss of classification accuracy.

In the library, recognizing the activity of the reader can better uncover the reading habit of the reader and make book management more convenient. In this study, we present the design and implementation of a reading activity recognition approach based on passive RFID tags. By collecting and analyzing the phase profiling distribution feature, our approach can trace the reader's trajectory, recognize which book is picked up, and detect the book misplacement. We give a detailed analysis of the factors that can affect phase profiling in theory and combine these factors with relevant activities. The proposed approach recognizes the activities based on the amplitude of the variation of phase profiling, so that the activities can be inferred in real time through the phase monitoring of tags. We then implement our approach with off-the-shelf RFID equipment, and the experiments show that our approach can achieve high accuracy and efficiency in activity recognition in a real-world situation. We conclude our work and further discuss the necessity of a personalized book recommendation system in future libraries.

Activity recognition from sensors is a classification problem over time-series data. Some research in the area utilize time and frequency domain handcrafted features that differ between datasets. Another categorically different approach is to use deep learning methods for feature learning. This paper explores a middle ground in which an off-the-shelf feature extractor is used to generate a large number of candidate time-domain features followed by a feature selector that was designed to reduce the bias toward specific classification techniques. Moreover, this paper advocates the use of features that are mostly insensitive to sensor orientation and show their applicability to the activity recognition problem. The proposed approach is evaluated using six different publicly available datasets collected under various conditions using different experimental protocols and shows comparable or higher accuracy than state-of-the-art methods on most datasets but usually using an order of magnitude fewer features.

In this study, toward the development of smartphone-based monitoring system for life logging, we collect over 1,400 hours of data by recording including both the outdoor and indoor daily activities of 19 subjects, under practical conditions with a smartphone and a small camera. We then construct a huge human activity database which consists of an environmental sound signal, triaxial acceleration signals and manually annotated activity tags. Using our constructed database, we evaluate the activity recognition performance of deep neural networks (DNNs), which have achieved great performance in various fields, and apply DNN-based adaptation techniques to improve the performance with only a small amount of subject-specific training data. We experimentally demonstrate that; 1) the use of multi-modal signal, including environmental sound and triaxial acceleration signals with a DNN is effective for the improvement of activity recognition performance, 2) the DNN can discriminate specified activities from a mixture of ambiguous activities, and 3) DNN-based adaptation methods are effective even if only a small amount of subject-specific training data is available.

Collective activity recognition plays an important role in high-level video analysis. Most current feature representations look at contextual information extracted from the behaviour of nearby people. Every person needs to be detected and his pose should be estimated. After extracting the feature, hierarchical graphical models are always employed to model the spatio-temporal patterns of individuals and their interactions, and so can not avoid complex preprocessing and inference operations. To overcome these drawbacks, we present a new feature representation method, called attribute-based spatio-temporal (AST) descriptor. First, two types of information, spatio-temporal (ST) features and attribute features, are exploited. Attribute-based features are manually specified. An attribute classifier is trained to model the relationship between the ST features and attribute-based features, according to which the attribute features are refreshed. Then, the ST features, attribute features and the relationship between the attributes are combined to form the AST descriptor. An objective classifier can be specified on the AST descriptor and the weight parameters of the classifier are used for recognition. Experiments on standard collective activity benchmark sets show the effectiveness of the proposed descriptor.

Activity recognition has recently been playing an important role in several research domains, especially within the healthcare system. It is important for physicians to know what their patients do in daily life. Nevertheless, existing research work has failed to adequately identify human activity because of the variety of human lifestyles. To address this shortcoming, we propose the high performance activity recognition framework by introducing a new user context and activity location in the activity log (AL2). In this paper, the user's context is comprised by context-aware infrastructure and human posture. We propose a context sensor network to collect information from the surrounding home environment. We also propose a range-based algorithm to classify human posture for combination with the traditional user's context. For recognition process, ontology-based activity recognition (OBAR) is developed. The ontology concept is the main approach that uses to define the semantic information and model human activity in OBAR. We also introduce a new activity log ontology, called AL2 for investigating activities that occur at the user's location at that time. Through experimental studies, the results reveal that the proposed context-aware activity recognition engine architecture can achieve an average accuracy of 96.60%.

With the development of personal electronic equipment, the use of a smartphone with a tri-axial accelerometer to detect human physical activity is becoming popular. In this paper, we propose a new feature based on FFT for activity recognition from tri-axial acceleration signals. To improve the classification performance, two fusion methods, minimal distance optimization (MDO) and variance contribution ranking (VCR), are proposed. The new proposed feature achieves a recognition rate of 92.41%, which outperforms six traditional time- or frequency-domain features. Furthermore, the proposed fusion methods effectively improve the recognition rates. In particular, the average accuracy based on class fusion VCR (CFVCR) is 97.01%, which results in an improvement in accuracy of 4.14% compared with the results without any fusion. Experiments confirm the effectiveness of the new proposed feature and fusion methods.

This paper proposes a method for using an accelerometer, microphone, and GPS in a mobile phone to recognize the movement of the user. Past attempts at identifying the movement associated with riding on a bicycle, train, bus or car and common human movements like standing still, walking or running have had problems with poor accuracy due to factors such as sudden changes in vibration or times when the vibrations resembled those for other types of movement. Moreover, previous methods have had problems with has the problem of high power consumption because of the sensor processing load. The proposed method aims to avoid these problems by estimating the reliability of the inference result, and by combining two inference modes to decrease the power consumption. Field trials demonstrate that our method achieves 90% or better average accuracy for the seven types of movement listed above. Shaka's power saving functionality enables us to extend the battery life of a mobile phone to over 100 hours while our estimation algorithm is running in the background. Furthermore, this paper uses experimental results to show the trade-off between accuracy and latency when estimating user activity.

An algorithm for the discrimination between human upstairs and downstairs using a tri-axial accelerometer is presented in this paper, which consists of vertical acceleration calibration, extraction of two kinds of features (Interquartile Range and Wavelet Energy), effective feature subset selection with the wrapper approach, and SVM classification. The proposed algorithm can recognize upstairs and downstairs with 95.64% average accuracy for different sensor locations, i.e. located on the subject's waist belt, in the trousers pocket, and in the shirt pocket. Even for the mixed data from all sensor locations, the average recognition accuracy can reach 94.84%. Experimental results have successfully validated the effectiveness of the proposed method.

In this letter, we propose a novel approach to human activity recognition. We present a class of features that are robust to the tilt of the attached sensor module and a state transition model suitable for HMM-based activity recognition. In addition, postprocessing techniques are applied to stabilize the recognition results. The proposed approach shows significant improvements in recognition experiments over a variety of human activity DB.

This paper describes a method for detecting multiple overlapping objects from a real-time video stream. Layered detection is based on two processes: pixel analysis and region analysis. Pixel analysis determines whether a pixel is stationary or transient by observing its intensity over time. Region analysis detects stationary regions of stationary pixels corresponding to stopped objects. These regions are registered as layers on the background image, and thus new moving objects passing through these layers can be detected. An important aspect of this work derives from the observation that legitimately moving objects in a scene tend to cause much faster intensity transitions than changes due to lighting, meteorological, and diurnal effects. The resulting system robustly detects objects at an outdoor surveillance site. For 8 hours of video evaluation, a detection rate of 92% was measured, which is higher than traditional background subtraction methods.

In this paper, a process is described for analysing the motion of a human target in a video stream. Moving targets are detected and their boundaries extracted. From these, a "star" skeleton is produced. Two motion cues are determined from this skeletonization: body posture, and cyclic motion of skeleton segments. These cues are used to determine human activities such as walking or running, and even potentially, the target's gait. Unlike other methods, this does not require an a priori human model, or a large number of "pixels on target". Furthermore, it is computationally inexpensive, and thus ideal for real-world video applications such as outdoor video surveillance.