This paper proposes a novel e-textile-based pressure sensor. Textile is common material in our life, used such items as sheets, seats, and clothing. If these are equipped with sensor functions, they can invisibly assist humans without significant lifestyle changes. Our suitable for mass production durable daily hard use cases. The woven weaving machines special manner its low-cost conductive fiber that does not costly materials, optical fiber. mechanism supported by the textile structure;...

E-textiles have come to be used instead of several types common equipment, such as bed-sheets, in some cases. An application using body pressure data collected through bed-sheet type sensors is the in-bed posture classification expected for ulcer prevention. Since a kind low-resolution image, Deep Neural Network (DNN) based algorithms seem suitable. However, it difficult collect enough use DNN this domain because number sleep postures obtained from one experiment small. For an example, 19...

We describe the results of human joint angle estimation with an e-textile-based stretch sensor. Joint is necessary to predict details posture and implement instruction prevent caregiver's injury in health medical care. In this study, we focused on elbow angle. installed e-textile sensors support maintain ideal setting a knitted shirt simulate daily-use setting. With support, correlation coefficient rooted-mean-square error degree arc were 0.99 5.73 for complete bends. shirt, same way,...

Multi-camera videos with abundant information and high flexibility are expected to be useful in a wide range of applications, such as surveillance systems, web lecture broadcasting, concerts sports viewing, etc. Viewers can enjoy high-presence viewing experience their own choosing by means virtual camera switching controlling interfaces. However, some viewers may feel annoyed continual manual viewpoint selection, especially when the number selectable viewpoints is relatively large. In order...

Many IMU-data-based HAR datasets have been proposed. However, integrating the use of those is difficult because each dataset measured by its unique sensor positions. Thus, we developed a garment device with 396 IMU sensors placed in grid approximately 7 cm spacing to enable matching data from various positions and integrated multiple databases. In this paper, best most suitable for classification were explored six basic activities eight daily expectation determining key creating standard...

Daily continuous spirometry is expected to detect lung disease initial symptoms that occur in daily contexts. Current medical spirometers are insufficient for such usage because they force uncomfortable conditions users, as pinching their noses with a clip and holding mouthpiece mouths. To solve this issue, we propose an e-textile-based wearable spirometer called SpiroVest estimates behavior from torso-girth movements. It does not require any interfere activity. In addition, our able...

Multiview videos can provide diverse information and high flexibility in enhancing the viewing experience. User-dependent automatic viewpoint recommendation is important for reducing user stress while selecting continually suitable favorable viewpoints. Existing personal methods have been developed by learning user»s records. These difficulty acquiring sufficient records practice. Moreover, they neglect importance of attribute information, such as personality, interest, experience level...

Accidents that are caused by "texting while walking" (TWW) continue to increase. Even though most people realize this activity is risky and should be avoided, more than a few engage in it. Since eradicating TWW appears difficult, using safe-walking support system an alternative reduce accidents during TWW. In paper, we first identify risk based on our preliminary experimental results, then propose system. For evaluation, compared the collision rates among four assessment algorithms three...

Daily continuous spirometry is expected to detect the initial symptoms of lung disease that appear in daily-life situations. Spiro Vest an e-Textile-based wearable spi-rometer suitable for daily because it does not require uncomfortable attachments, such as nose-pinching clips and mouthpieces used by typical medical spirometers. However, several adaptation issues changes context must still be resolved. This paper discusses two these remaining issues: sweat diet. The due diet significantly...

On-demand, skill-level self-checks are required to establish effective training, and this is the same for metal-filing. However, such a training system has not been established yet metal-filing because current check depends on subjective estimation by experts. Such composed of various complex viewpoints, its mechanisms cannot be represented in language they generated supported experience That why sensor-based, on-demand systems faithfully imitate or replace checks To solve problem, we...

This paper provides a novel method for building non-task-oriented dialogue agents such as chatbots. The dialogue agent constructed using our automatically selects suitable utterance depending on context from set of candidate utterances prepared in advance. To realize automatic selection, we rank the candidate order suitability by application machine learning algorithm. We employed both right and wrong data to learn relative the utterances. Additionally, provide a low-cost...

SUMMARY As a steady new network communications tool, social media have reached global proportions. This phenomenon has had an impact on societies all over the world. Above all, people provide information to Twitter and Facebook daily basis. result, vast amounts of data exist Facebook, we can expect gather useful from these services. was changed by 2011 Tohoku Earthquake. Further, greatly contributed diffusion information. For example, many users checked safety their friends or family....

This paper presents a method to evaluate person's skill level for metal filing. Metal filing by expert engineers is an important manufacturing that supports basic areas of industry, although most sequences are already automated with industrial robots.

Many deep learning (DL) models have been proposed for the IMU (inertial measurement unit) based HAR (human activity recognition) domain. However, combinations of manually designed time series features (TSFs) and traditional machine (ML) often continue to perform well. It is not rare that among TSFs DL show better performance than DL-only approaches. Those facts mean potential outperform automatically generated using neural networks (DNNs). a problem: their performances are only good if...

Elderly people using wheelchairs have higher risks of developing pressure ulcers and falling out the chairs. Caregivers to monitor postures such elderly correct dangerous posture. It is high-workload care because it must be conducted throughout day. Therefore, our project aims reduce workload by ICT-based posture monitoring with smart wheelchairs. In this paper, we discuss line-oriented sensing point reduction for non-categorical quantitative score based sitting evaluation. experiment, even...

E-Textiles using fabric sensors have been studied well for respiration monitoring in recent years; they can estimate respiratory rates and patterns. However, studies of volume flow estimation remain unestablished, although are necessary the inspection chronic obstructive pulmonary disease (COPD). In this paper, we introduce a new stretchable textile sensor examine how to calculate volume. The stretch up about 150% (e.g., 13 20 cm). be detected from electronic potential changes between...

Our project is estimating and reducing the risk of pressure ulcers using bed-sheet-type textile sensors. However, body-pressure dispersion cushions interfere with accurate sensing cause such problems as indistinct changes, blurred body shape, unmeasurable changes. To solve these issues, we implement a garment-type sensor that measures pressures are directly added to human skin. successfully detected changes overlooked by sensors, avoided shape vagueness, measured increase depending on part...

Multi-camera videos with abundant information and high flexibility are useful in a wide range of applications, such as surveillance systems, web lectures, news broadcasting, concerts sports viewing. Viewers can enjoy an enhanced viewing experience by choosing their own viewpoint through interfaces. However, some viewers may feel annoyed the need for continual manual selection, especially when number selectable viewpoints is relatively large. In order to solve this issue, we propose automatic...

The displacement of sensors integrated into smart garments for human behavior sensing increases estimation error. To resolve this issue, we propose reducing error by intermittent adjustment using posturestable daily actions. We introduce approach and describe our evaluation the actions drinking from vessels holding a doorknob to estimate elbow joint angle. two provided suitable characteristics proposed adjustment. results showed 103.6 degrees arc reduction in best case scenario.

Sleeping posture estimation including joint positions is important for identifying pressure ulcer risk. Since there are occlusion and privacy problems with camera images, we have been studying 2D position from sleeping images. However, the does not reveal 3d relationships among body parts, such as crossing legs. Thus, toward fine-grained activity recognition, 3D required. To study it, extend joint-position data of SLP dataset then tried to estimate them high accuracy. In this paper,...