1 Introduction

With the advancement in convergence technology, interactive communication using context awareness computing environment and various devices became available and great attention is being paid to the studies on health management. As there is an increase in chronic disease followed by the change in life habits, increase in quality of life, and lengthening of life expectancy and as the health management is becoming an important issue in medical service industry, the necessity for continuous management and health improvement has been promoted instead of previous approach to treat the disease. Particularly, dramatic aging followed by the lengthening of life expectancy is estimated to cause an increase in chronic diseases and complications and thus increase in total medical cost [2]. Companies have started to join the trend of health management service to get accustomed to such environment and make bold investment to BT–IT convergence project for health management services. They are not only promoting test-projects in relation with on/offline health management system but also promoting the nationwide health management service for chronic diseases as a business. Representative U-health services provide health management through various networks including WebMD Health Manager PHRFootnote 1, Microsoft HealthVaultFootnote 2, Google Health, Patient AllyFootnote 3, and others. There is a necessity for custom-made health management service which automatically provides fundamentally differentiated information by approaching health management of chronic disease as the issues of family and society rather than putting emphasis solely on individual level just like in the past and applying it to overall problem resolution [3–6].

With overall environment change in medical service industry such as the increase in medical cost of chronic disease patients and advancement in convergence technology of BT and IT, there is a necessity for conversion from medical institute centered medical service system to that of medical service consumer centered. Namely, it shall be converted to a format in which chronic disease patients actively manage and utilize personal health record (PHR) for themselves. The means to actively intervene in health management and treatment process and participate in decision making process shall be provided for chronic disease patients who have been passively corresponding to medical institute centered medical service provision system. Also, the service shall be provided so that medical service consumer himself receives, manages, maintains, and monitors the information relevant to his health. Integrated model is necessary in order to achieve clear communication and interactive operation between interested parties so that chronic disease patients can easily use the service in social life. The trait of integrated model is the interworking of medical information that exists outside the hospital information system of medical institutes through network and the environment in which patient himself can manage his own PHR by receiving health related information from various medical information providers. Through such model, health related information and patterns of bio expertise can be received at anytime and anywhere instead of receiving them from particular places only in the past [5–9]. In short, effective health management and improvement will be available with the composition of practical services necessary for chronic disease patients.

This study is structured as follows: Section 2 examines other related studies of U-health service systems and healthcare assistant applications for the chronic disease. Section 3 deals with the context motion tracking of chronic disease patients. Section 4 assesses the development of emergency situation monitoring service using the context motion tracking. This research conclusion is explained in Sect. 5.

2 Related works

2.1 U-health service system for chronic disease

The importance of U-health service for chronic disease patients are gradually increasing with the change in dietary life of people, decrease in amount of physical activities, and entry to aging society. With dramatic increase in chronic disease patients due to illness, there also is an increase in occurrence of complications including hypertension, diabetes, and others and furthermore indifferent management by patient himself may bring about severe pain and unrecoverable state [10]. Therefore, there is a necessity for convergence technology combined of BT and IT which can be managed at anytime and anywhere. Context awareness service defines the subject related to patient information as the context and computer provides custom-made service by recognizing the time and whereabouts of patients [11].

There exists various techniques including big data, ontology, inference, training, data mining, decision tree, entropy, information filtering, artificial intelligent, and others to provide context awareness based service and provision of proper health improvement service is available with above techniques [12, 13]. More attention is pay to the provision of context awareness computing based medical service and context awareness computing that can correspond to various situations of patients can increase the satisfaction of service with intelligent management of contextual information [14]. Also, the provision of active service is available since the purpose of utilizing U-health environment lies in providing high quality health service through inference of all contextual data related to patient [7–9, 15, 16].

Previously developed remote monitoring system for daily activities of single residence elderly [17] can prevent the problems that may occur by monitoring problems which occur due to lowering of vision, low activity, and decrease in memory in daily lives. As a previous single residence elder assistance system using wireless RFID glove, it is composed of wearable RFID system, home health gateway [1], and server system. Wearable RFID system transmits the information of RFID tag attached object to gateway by using wireless sensor network and RFID reader. Gateway informs elderly the information on object with video and audio from RFID tag information. A single residence elder protection system using sensor network grasps real-time the behavior of single residence elderly and provides such information to external institute, family, and relatives through wire and wireless connection and it enables immediate measures at emergency. Also, bio information of single residence elderly can be acquired with attachment of bio sensors to living space of single residence elderly and judgment on context is available real-time [18–20]. It is urgent to come up with integrated precautions as dramatic increase in medical treatment is estimated due to chronic diseases, physical hypofunction, and disabilities with entry to the aging society. In order to prevent the occurrence of geriatric illness, it is necessary to detect and manage in advance the patients with metabolic diseases including hypertension, diabetes, hyperlipidemia, cataract, glaucoma, arthritis, renal failure, cerebrovascular accident, myocardial infarction, hepatitis, and others which are risk factors of chronic disease.

2.2 Healthcare assistant application of chronic disease

There is an increase in prevalence and occurrence of chronic degenerative diseases which are geriatric disease due to the aging of population and increase in average life expectancy. Chronic degenerative diseases lower the quality of life for elderly population not only by limiting physical activities but also accompanying physical and psychological pain and others. Furthermore, it is main cause of aggravating the budget burden of national medical cost including increase in expense for national health insurance and others.

NameTag previously developed to overcome dementia is an application manufactured by Facial Network and it recognizes the face by using camera and displays the profile that matches it to the screen [21]. Figure 1 displays the screen of NameTag (http://www.nametag.ws). When the profile of family members is saved through this, the patient can recognize the face of family members even with the symptoms of dementia. The scope of information to open to public can be designated at personal setting for privacy protection and another profile to be displayed at non-working hours can be saved. Dementia is one of geriatric illnesses and it is a disease with highest increase rate for the last 5 years. Studies on brain are actively being conducted to lower the prevalence of dementia.

Fig. 1
figure 1

Screen of NameTag (http://www.nametag.ws/)

Full size image

Med Helper is personal medical recording application and prescription can be recorded in details. Important information such as the date of prescription reception, period, name of medication, administration interval, caution, information of chronic disease, and others are recorded and it enables proper administration of drug through push alert function [22]. In case of long-term administration, the prescription management is available by recording the amount of prescription and administration. As a medical guideline service for prescription, it automatically sends the message based on the guidelines to prevent the side effects of drug. Guardian can manage several patients at once with the use of multi profile function. Recorded data can be created into the document and printed out and it supports iOS and Android operating system. Figure 2 shows the application of Med Helper for healthcare assistant (www.medhelperapp.com).

Fig. 2
figure 2

Application of Med Helper for healthcare assistant (http://www.medhelperapp.com/)

Full size image

3 Context motion tracking of chronic disease patients

3.1 System overview

The emergency situation monitoring system service observes the movement of user by using video transmitted from several cameras to conduct context motion tracking. Also, selective use of home health gateway using IEEE 11073 manager parsing [23], and wire/wireless communications is available for data transmission. The emergency situation alarm system research of [1, 2] is described more detail. In here, home health gateway refers to the standardized technology which enables smooth data exchange and more effective processing and management of data reuse, analysis, search, and statistical information [24–28]. Information necessary for chronic disease management is provided through determination on current status of patients and analysis on life habits with contextual information collected through the system. Context motion tracking provides emergency situation service accordingly with alert and symptom level in case of symptom occurrence through measured results and analysis.

The system determines the occurrence of symptom as semi-emergency situation when the symptoms are detected from behavior patterns at active movement time and the movement is induced by executing sound data “Move” through speaker in unit of 10 s for total 5 times. When the result of context motion tracking is weak or no response, it is determined as emergency situation and video is transmitted for 10 s and message to notify emergency situation are sent to family/guardian and medical center. The context motion tracking using the inference engine is visually very intuitive thus it can detect the smallest movement and it is proper technology to detect the emergency situation. Figure 3 shows the emergency situation monitoring service using context motion tracking of chronic disease patients.

Fig. 3
figure 3

Overview of proposed emergency situation monitoring service

Full size image

This system constitutes normal household composed of living room and other rooms where chronic disease patients lead daily lives as the context. In order to collect contextual information, cameras, speakers, and measurement sensors are installed to living room and other rooms as illustrated in Fig. 4. The measurement sensors include temperature, humidity, illumination, and motion sensors. This is due to the fact that chronic disease patients can move around many places within house including the living room thus contextual information of user is created by measuring real-time the indoor temperature, humidity, and illumination. In here, the measurement sensors are operated with wire/wireless connection thus it has advantage of easy installation and use. Life log based contextual information can be collected through extended use of various sensors. The cameras and motion sensors are necessary to conduct context motion tracking of chronic disease patients. Also, the speakers are necessary since it needs to deliver “Move” message in case there is no movement detected for certain period of time.

Fig. 4
figure 4

Configuration of hardware installed in smart health home

Full size image

The context awareness which can correspond to various situations of chronic disease patients can increase the satisfaction of smart health service with the management of intelligent contextual information. The health status and life pattern can be analyzed with semantic inference engine which uses contextual information [26–28]. Also, the utilization of contextual information can provide custom-made service through inference of all contextual data related to chronic disease patients. Semantic inference engine for context awareness conducts active and intelligent analysis on health status and life patterns. Since it properly corresponds in extraordinary situations, it provides proper service environment at emergency situation or symptom occurrence in the smart health service. As illustrated above, this system saves video data of cameras, temperature, humidity, illumination, and motion data for motion recognition and tracking transmitted with wire/wireless connection from each room including the living room to contextual information database. After completing the ontology based context awareness with contextual information database, it provides real-time display through user interface. Then, the administrator can properly provide remote health management and medical service with the use of feedback function [29–34].

3.2 Gesture context awareness using motion history image

Motion history image (MHI) [35, 36] is used in this study for motion recognition and continuous tracking from video. MHI, an appearance based gesture recognition method [24, 25, 37], refers to motion history image which expresses pixels of most recently moved domain in brighter value with the use of pixel intensity and it is basic method to express the behavior. Basic MHI \(\text{ H }_{\tau } (\text{ x },\text{ y },\text{ t })\) can be calculated using update function \({\phi }(\text{ x },\text{ y },\text{ t })\) as illustrated in Eq. (1).

$$\begin{aligned} \mathrm{H}_{\tau }\mathrm{(x,y,t)}\!=\!\! \left\{ \! {\begin{array}{ll} {{\tau }} &{} \quad \mathrm{if}\;\varphi \mathrm{(x,y,t)} \!=\! 1 \\ {\text {max(0,H}}_{\tau } \mathrm{(x,y,t \!-\! 1)} \!-\! \delta ) &{}\quad \mathrm{otherwise} \\ \end{array} } \!\right\} \end{aligned}$$
(1)

In here, \((x,y)\) indicates the position and \(t\) indicates the time. Therefore, \(\varphi (x,y,t)\) informs the existence or motion of object from current video image. \(t\) determines time range of movement in aspect of frame and it indicates the maximum length of time decay. \(\delta \) is decay parameter. This update function \(\varphi ( \mathrm{x,y,t})\) is called for every new video frame analyzed in the sequence. The calculation result of Eq. (1) is scalar-valued image of which more recently moved pixels are expressed brighter and it is same for vice versa. In here, some possible image processing techniques for defining this update function \(\varphi ( \mathrm{x,y,t})\) include background subtraction, image differencing, and optical flow. Generally, MHI is created from binary coded image acquired from frame subtraction which uses threshold \(\xi \) as illustrated in Eq. (2) [24].

$$\begin{aligned} \varphi ({\text {x,y,t}}) = \left\{ {\begin{array}{ll} 1 &{} \quad \mathrm{if}\;{\text {D(x,y,t)}} \ge \xi \\ 0 &{} \quad \mathrm{otherwise} \\ \end{array} } \right\} \end{aligned}$$
(2)

In here, \(\text{ D }(x,y,t)\) is defined as difference distance \(\Delta \) same as Eq. (3).

$$\begin{aligned} \text{ D }( \mathrm{x,y,t})=\left| {\text{ I }( \mathrm{x,y,t})-\text{ I }(\mathrm{x,y,t}\pm \Delta )} \right| \end{aligned}$$
(3)

In here, \(\text{ I }(x,y,t)\) signifies the intensity value of pixel location with coordinate \((x,y)\) from no. t frame of image sequence. Ultimately, MHI template becomes \(\text{ H }_{\tau } (x,y,\tau )\) and threshold \(\xi \) of Eq. (2) is set as 100 in this study to be used for the experiment as the threshold \(\xi \) of less than 100 signifies the smallest movement of object. Therefore, the threshold \(\xi \) was set as 100 so that it would exceed the range of minute movement.

3.3 Semantic inference engine for context awareness

Inference is defined as an act of making an assumption about a certain fact on the basis of several relevant evidences and context. A semantic inference engine is a computer program to perform such an assumption process. It has the storage of processing rules to be applied to fact relevances or situational conditions previously clarified at a specific area of knowledge base, so that it can calculate new facts or make conclusions by using the sets of the previously stored facts and rules or output a list of actions suitable to the present query or situation. The data to be input to the semantic inference engine may be a set of attribute values that describe a situation or some properties of an object or event associated with the situation. The output data depend on the purpose or function of a certain system. For example, if the input data to an obesity management expert system of a medical domain is the recent two week weight change graph of a particular individual, the output data may be information on the next several days’ food intake on the basis of the input data [38–40].

Figure 5 shows the process of context awareness for health status and life patterns. Generally, the knowledge base used by the semantic inference engine is created by knowledge engineers. The knowledge is what is reproduced with a special code that can be interpreted by a program which maps between the problems that can occur at a domain and their corresponding solutions. The reproduction process can be interpreted by learning. Learning of knowledge is divided into two methods. The one is that knowledge engineers make direct inputs, and the other is that the algorithms that analyze data to be produced in the domain make an automatic acquisition. The tools of encoding and learning the domain knowledge and the semantic inference engine of processing the input data on the basis of the domain knowledge are called an expert system together. The semantic inference engine is classified depending on key algorithms and knowledge base construction methodologies. On the basis of the \(<\)If-Then\(>\) generation rule, the rule-based semantic inference engine outputs a list of actions to be taken by systems or users or conclusions to be made by the current context awareness. Meanwhile, the semantic inference engine makes an inference on the basis of the domain knowledge modeling ontology knowledge base [34].

Fig. 5
figure 5

Process of context awareness for health status and life patterns

Full size image

The condition and context awareness of chronic diseases was defined with the use of semantic inference engine and attribute information that can represent disease information was used based on physical examination survey data of the Korea National Health and Nutrition Examination Survey (KNHNES) [41]. The KNHNES was performed by the Korea Center for Disease Control and Prevention, the Ministry of Health and Welfare (http://www.mw.go.kr/). The KNHNES includes reliable statistical data on national health statistics, health related awareness and behaviors, and the current state of food and nutrient intakes at all levels (city/province/county levels). Those basic data are essential in establishing and assessing national healthcare policy including setting goals, evaluating the National Health Promotion Plan and developing health programs [15]. Physical examination survey available for acquisition of attribute information in regards to chronic diseases include blood pressure and pulse measurement, exercise, smoking habits, blood and urine test, BMI examination, and body mass index. For the attribute information determination of chronic disease patients, rule-based algorithm was utilized to conduct preconditioning process based on selected condition. Preconditioning is divided into a process to remove omitted items and unnecessary items and a normalization process based on rule-based algorithm condition. Higher appearance frequency of the same attribute information among patients signifies more close relation between patients in aspect of contents [6]. By conducting the classification based on test value or existence of health related items of physical inspection items, the provision of health and disease information is determined and afterwards smart health service is proposed when there are items associated with each category. Proper service shall be proposed after examining whether risk factors including body mass index, albuminuria, cholesterol, smoking, blood sugar, exercise, and drinking have been checked or not. The provision of smart health service with semantic inference engine enables the disease management, improvement in life habits, and enhancement in self-management ability.

4 Development of emergency situation monitoring service using context motion tracking

4.1 User interface

The composition of this system is based on living space of chronic disease patients thus there is necessity for a viewer which can display real-time at least 3 camera image data. For the real-time display of context motion tracking video which uses camera image data, the same number of viewers as cameras is necessary. Also, since it is targeting chronic disease patients, the motion capture video is created and its result is displayed with camera which captures the movement as its subject when the user moves around the living room and other rooms. Therefore, it tracks the movement of chronic disease patients and display real-time the size of movement with vertical bar graph composed of horizontal gradations. A bar graph to display real-time the size of movement was placed between camera image viewer and motion tracking video viewer. Table 1 shows the proposed system composition and interoperating data. The proposed system was designed to use portable devices through the design for web browsers and mobile devices.

Table 1 System composition and interoperating data
Full size table

Figure 6 shows the user interface of this system for motion tracking test scene. The motion tracking results of chronic disease patients is very intuitive, visually as shown in the figure, because a little bit of movement can be detected. Therefore, motion tracking techniques is good techniques as the criteria for detection of emergency situation of chronic disease patients.

Fig. 6
figure 6

Initial screen and user interface

Full size image

The placement of play/launch frame composed of button type menu to operate/stop this system or save each of video as an image file was conducted. Also, the system was equipped with sensor information frame to display the temperature, humidity, illumination, and motion data of living room and other rooms and calendar frame was placed as well for daily check-up and recording of the date. Since this system is operated 24 h every day in order to track the movement of chronic disease patients, it was equipped with motion size per time frame to view the size of movement per hour by using data of yesterday. Also, it possesses as the contextual information the motion ratio per room frame to display the ratio for movement of chronic disease patients at which place and what duration with living room and all other rooms as its subject. The system also include real-time motion tracker frame to track real-time the movement of user with living space as its subject and display it as a graph. It conducts context awareness through semantic inference engine based on video transmitted real-time from cameras installed to living room and each room. It is composed of contact information frame to input necessary information for transmission of PHR regarding chronic disease patients and emergency situation information upon the emergency situation, symptoms, and peculiar behavior patterns. The frame for above measures was separately formed as illustrated in Fig. 7. For the convenience in management at home healthcare management center, medical center, and others, PHR on chronic disease patients includes photo, ID, name, password, gender, birthday, home tel., mobile phone, address, GPS, and case history. In here, the case history is for the reference of emergency room regarding medical history, allergic reaction, EMR, EHR, and PHR on previous medical history and current chronic disease is also recorded. Moreover, GPS is used as contextual information for air ambulance and ambulance to accurately grasp the location of chronic disease patients together with the address.

Fig. 7
figure 7

Customer information and emergency contacts

Full size image

In regards to the emergency contact information in case of emergency situation occurrence, it enabled the input of phone numbers for guardian, police station, ambulance 119, hospital, emergency room, social worker, and others. In short, wire and wireless emergency situation monitoring service has been developed using context awareness so that one can reach medical center very promptly and accurately in case of emergency situation occurrence.

4.2 Implementation results

Figure 8 is an active screen shot which displays the transmitted screen of currently operating system which has been developed in this study. Namely, 3 videos transmitted from 3 cameras and motion tracking video can be viewed real-time and the progress of motion tracking from the camera video which captures the movement of chronic disease patients can also be examined. In here, Sensor Information, Motion Size per Time, Motion Ratio per Room, and context Motion Tracker are developed considering the user convenience by expressing it as bar graph, line graph, and pie chart using Chart Director [42]. Sensor information frame displays the temperature, humidity, illumination, and motion data measured from living room and other rooms in a bar graph. Motion Size per Time frame displays the movement quantity of chronic disease patients during 24 h as a line graph so that it can be viewed at a glance. Also, Motion Ratio per Room frame displays the movement quantity of chronic disease patients in living space during 24 h as a pie graph so that the ratio of each movement can be viewed at a glance. It assists the chronic disease patients to change into proper life habit in aspect of exercise, dietary habit, and health management appropriate for current situation. Systematic self-health management and life patterns can be analyzed from life log which was collected through daily/weekly/monthly statistical data. Life patterns and psychological state measurement are available through life log based motion detection and pattern of expertise based on semantic context inference can be determined with long-term observation of change in particular motion.

Fig. 8
figure 8

Active screen shot of developing system

Full size image

Context motion tracker frame expresses the progress of real-time motion tracking from camera installed to living space as a line graph. Namely, it displays not only the size of motion from video transmitted from camera in numerical data but also as a line graph so that the size of movement occurrence at whichever camera or living space can be well examined. Although the experiment on this system was basically conducted with living room and other rooms as its subject, the expansion of system is available with the attachment of several cameras & speakers and temperature, humidity, illumination, and motion sensors accordingly with the structure of living space the chronic disease patients reside in.

5 Conclusions

With the change in lifestyle and aging, the approach for chronic disease is changing from previous disease treatment centered into health improvement and management centered. Also, with the advent of aging society due to dramatic economic growth and advancement in medical techniques followed by industrialization and advancement in science and, there is a trend of increase in chronic diseases rather than acute diseases. Chronic diseases require the support of continuous health management and expanded service which provides patients with management and check-up schedule based on behavior of patients different from treatment centered medical service. In advanced countries, great attention is paid to the increase in economic power of the elderly which exceeds its population ratio followed by dramatic increase in the aged population, users of health management service at charge including hospice, nursing, and others. With the increase in number of chronic disease patients, U-health service is changing from the treatment and management of certain disease to behavior modification based preventive service for the service target. Previous U-health service has limitation in prevention and management of chronic disease such as metabolic syndrome and others as it only provides the service through remote physical examination and counseling, monitoring of bio information, and others. In this paper, we proposed the emergency situation monitoring service using context motion tracking for chronic disease patients. We proposed the emergency situation monitoring system that can detect emergency situation by transmit message including camera images and context information quickly to the people around them if an emergency situation occurs to chronic disease patients. This developing system is used the context motion tracking techniques based on transferred images from multiple cameras.

The emergency situation monitoring system was designed and implemented targeting the living space where chronic disease patients reside in. Cameras & speakers, and temperature, humidity, illumination, and motion sensors were installed to living room and each room and they were connected to developed system through wire and wireless communication. The system transmits the data real-time to home healthcare management center and medical center through home health gateway with the use of wire and wireless communication device in case of emergency situation occurrence. Also, this system automatically and immediately contacts previously registered guardians, family, social workers, and acquaintances in case of emergence situation occurrence in living space. Therefore, it provides health management and emergency situation monitoring service which can be used at anytime and anywhere and it can be used easily with simple operation considering the user convenience. In short, wide-ranging wire and wireless emergency situation service has been developed so that chronic disease patients can reach medical center very promptly and accurately in case of emergency situation occurrence.

Studies on interactive user interface shall be conducted in the future considering the diversity in age and the verification on efficiency of this developing system will be conducted through assessment and analysis on health maintenance and improvement in diverse users.