A local orientation descriptor (LOD) for nutrition analysis by quantity estimation is proposed. By observing nutrition properties, a texture-based LOD is designed to extract discriminant information, frequency and length among food items. Prior to classification, food detection is a challenging problem due to significant variety of backgrounds and containers. Thus, three food region detectors are designed in this study. A detector that employs a modified salient object detection algorithm using prior background knowledge provides promising segmentation results for non-uniform backgrounds. Integrating gradient information to construct graph weights yields more precise segmentation results. In addition, nutrition quantity is estimated using coins as reference objects. Three types of features, normalized colour, density, and symmetry properties are extracted for coin classification. Experimental results show that the proposed LOD outperforms existing object recognition features.

Multiscale entropy (MSE) is a measurement for quantifying the randomness of a sequence of data. Recently, it has been proven to be the most effective way to analyze the complexity of physiological signals in biomedicine and other fields. The implementation of MSE is computationally expensive because it considers multiple complexities of several data sequences with multiple scales and the computation of entropy for each data sequence is time consuming. A large number of original observed data for computing MSE is necessary because the number of data reduces when the scale increases. We must confirm that the data for each scale is enough to robustly obtain entropy. The large data problem makes MSE difficult for online application of processing sequential data. This paper provides a new online MSE computation method to improve computation efficiency of MSE. Furthermore, we apply distribution statistics in online MSE computation procedure to reduce the storage space of a system. In addition to segmenting the original data sequence into several non-overlapped sliding windows to reduce the data amount of each computation, different kinds of metadata are defined for metadata updating algorithm to accelerate computation time and save storage space. Experiments analyses with electrocardiogram (ECG) revealed that the proposed MSE estimation methods achieved significant improvement of more than 15 times faster than the conventional method for N= 60,000. Moreover, the proposed method only uses about 1/500,000 storage space compared to the conventional method.

Mining useful information and helpful knowledge from large databases has evolved into an important research area in recent years. Among the classes of knowledge derived, finding sequential patterns in temporal transaction databases is very important since it can help model customer behavior. In the past, researchers usually assumed databases were static to simplify data-mining problems. In real-world applications, new transactions may be added into databases frequently. Designing an efficient and effective mining algorithm that can maintain sequential patterns as a database grows is thus important. In this paper, we propose a novel incremental mining algorithm for maintaining sequential patterns based on the concept of pre-large sequences to reduce the need for rescanning original databases. Pre-large sequences are defined by a lower support threshold and an upper support threshold that act as gaps to avoid the movements of sequences directly from large to small and vice versa. The proposed algorithm does not require rescanning original databases until the accumulative amount of newly added customer sequences exceeds a safety bound, which depends on database size. Thus, as databases grow larger, the numbers of new transactions allowed before database rescanning is required also grow. The proposed approach thus becomes increasingly efficient as databases grow.

For the aging population and for people with dominant chronic diseases, countries all over the world are promoting an "Aging in Place" program with its primary focus on the implementation of telecare. In 2009, Taiwan held a "Health Care Value-Added Platinum Program" with the goal of promoting the development of "Telecare" services by integrating medical treatment, healthcare, information communication, medical equipments and materials and by linking related cross-discipline professions to enable people to familiarize themselves with preventive healthcare services offered in their household and community environments. In addition, this program can be utilized to effectively provide diversified healthcare service benefitting society as a whole. This study aims to promote a diversified telecare service network in Taiwan's household and community environments, establish telecare information platforms, build an internal network of various healthcare service modes, standardize externally interfacing telecare information networks, effectively utilize related healthcare service resources, and complete reasonable service resource links forming an up-to-date health information exchange network. To this end, the telecare information platform based on service oriented architecture (SOA) is designed to promote an open telecare information interface and sharing environment to assist in such tasks as developing healthcare information exchange services, integrating service resources among various different healthcare service modes, accessing externally complex community affairs information, supporting remote physiological information transmissions, and providing diversified remote innovative services. Information system architecture and system monitoring indices of various types of healthcare service modes are used for system integrations for future development and/or expansions.

This paper presents a guideline-driven healthcare service platform for smart homes. The clinical guidelines are usually established by medical experts according to the symptoms of diseases. Based on the biological and ambient information detected by sensors, the guideline-driven system makes the appropriate care decisions and then takes actions. In case of emergency, the platform uses the short message system to inform the relevant units immediately. The proposed service platform can therefore reduce the burdens of care and unexpected events in an effective way; therefore it enhances the quality of family care.

Mining knowledge from large databases has become a critical task for organizations. Managers commonly use the obtained sequential patterns to make decisions. In the past, databases were usually assumed to be static. In real-world applications, however, transactions may be updated. In this paper, a maintenance algorithm for rapidly updating sequential patterns for real-time decision making is proposed. The proposed algorithm utilizes previously discovered large sequences in the maintenance process, thus greatly reducing the number of database rescans and improving performance. Experimental results verify the performance of the proposed approach. The proposed algorithm provides real-time knowledge that can be used for decision making.

Modification of records in databases is common in real-world applications. Developing an efficient and effective mining algorithm to maintain discovered information as the records in a database are updated is thus quite important in the field of data mining. Although association rules for modification of records can be maintained by using deletion and insertion procedures, this requires twice the computation time needed for a single procedure. In this paper, we present a new modification algorithm to resolve this issue. The concept of pre-large itemsets is used to reduce the need for rescanning original databases and to save maintenance costs. The proposed algorithm does not require rescanning of original databases until a specified number of records have been modified. If the database is large, then the number of modified records allowed will also be large. This characteristic is especially useful for real-world applications.

In this study we attempt to quantify the correlations between time asymmetric index (TAI) and physiological/ pathological measurements in order to predict the status of cardiac autonomic neuropathy (CAN) in patients with end-stage renal disease. The TAI can explore non-equilibrium dynamics that conventional time and frequency domain indices fail to consider. For comparing the effectiveness and applicability, traditional time-domain and frequency-domain heart rate variability (HRV) indices were also evaluated. To validate the use of TAI, this study investigates 65 DM patients (22 men and 43 women) and divides the data into two subgroups by HbAlc. The HRV indices such as PNN50 and nLF exhibit a positive correlation with CTR in HbAlc+ subgroup, while SDANN and nLHR show a positive correlation between P and Caphos, respectively. In particular, TAI exhibits a significant negative correlation with TNF-alpha, HbAlc, Caphos, and GLUAC in HbAlc+ subgroup, which overcomes the limitation that a single index can be only correlated to one measurement for conventional time-domain and frequency-domain HRV indices. These results manifest that TAI is an integrated HRV index that contains more information than traditional HRV analysis. Moreover, TAI provides a useful and quantified method in evaluation of the cardiac autonomic neuropathy in patients with ESRD. A large cohort study with TAI analysis is in need for potential cardiovascular outcome prediction.

Keeping a healthy and balanced diet has long been a critical issue for a person wanting to stay fit and energetic in her/his daily life. We can always turn to a dietitian (or a nutritionist) for professional diet suggestions if necessary. However, we cannot have a dietitian staying with us all the time, which renders daily nutrition control very challenging. Therefore, it is desirable for each individual to receive handy and informative diet suggestions whenever necessary. In this work, we propose a Context-aware Personal Diet Suggestion System (CPDSS) which tries to maximize an aggregated health utility function and provides useful diet suggestions according to some contextual information. In order to increase the practicality of the pro- posed system, we have integrated the CPDSS with an everyday appliance–a smart refrigerator–so that we can readily access the suggestions about one’s diet, receive instant context-aware reminders while preparing foods, and keep long-term diet history to extract more useful pat- terns to caregivers to provide suggestions for health improvement.

Current trends suggest that the Pervasive systems should be adaptive and flexible. Many Message-Oriented Pervasive systems have been proposed to support on-line self-composition of services in reaction to changing contexts. However, most existing service composition mechanisms either do not consider user preferences or use proprietary typing systems. In this paper, we propose an extensible ontology-based autonomous service composition framework for Message-Oriented Pervasive systems. System developers can reconfigure this framework according to their application domains by overriding default hook functions. Experimental results show that the superiority 1 of the proposed approach become obvious along with increasing number of nodes.