Monitoring the activities of daily living (ADLs) and detection deviations from previous patterns is crucial to assessing ability an elderly person live independently in their community early upcoming critical situations. "Aging place" for one key element ambient assisted (AAL) technologies.

In this paper, we present the scientific outcomes of 2017 Data Fusion Contest organized by Image Analysis and Technical Committee IEEE Geoscience Remote Sensing Society. The was aimed at addressing problem local climate zones classification based on a multitemporal multimodal dataset, including image (Landsat 8 Sentinel-2) vector data (from OpenStreetMap). competition, separate geographical locations for training testing proposed solution, models that were accurate (assessed accuracy metrics...

This paper presents an end-to-end system for automatic local climate zones classification of various types urban environment. For that we perform fusion multispectral images from Landsat-8 and Sentinel-2 satellites with site description extracted OpenStreetMap layers. The proposed approach is based on a multi-level ensemble scheme combines Convolutional Neural Networks, Random Forests Gradient Boosting Machines.

Support Vector Machines (SVMs) are considered to be one of the most powerful classification tools, widely used in many applications. However, numerous scenarios classes not equally represented and predictive performance SVMs on such data can drop dramatically. Different methods have been proposed address moderate class imbalance issues, but there few that successful at detecting minority while also keeping high accuracy, especially when applied datasets with significant level imbalance. In...

In this paper, we present an ensemble-based classification approach for urban land use and cover based on multispectral LiDAR, hyperspectral very high resolution RGB data. The has been evaluated the data set provided IEEE GRSS 2018 Data Fusion Contest organized by IADF technical committee proven to have a operational performance, being able distinguish between different grass-, building- street-types among other classes like water, railways parking lots as well non-typical cars, trains,...

Data fusion (DF) from multiple heterogeneous sources is a typical task for many multisensor applications including remote sensing classification problems. Multiple classifier systems (MCS) provide natural way to solve DF on the decision level by training individual classifiers separately its own data source and then combine their outputs. In this paper, we consider dynamic selection (DS) framework select fuse competent of MCS. For this, propose competence estimation method improve...

Consensus clustering, also known as clustering ensembles is a technique that combines multiple solutions to obtain stable, accurate and novel results. Over the last years several consensus approaches were proposed addressing practical problems with different degrees of success. In this paper, we consider data fragments elements cluster ensemble framework. We propose new dissimilarity measure on build function allows handling large scale while not compromising accuracy. evaluate our number...

In this paper, we study the possibilities of applying theory interval systems linear algebraic equations to solve problems mathematical modeling processes using experimental data. It is assumed that simulated process described by output variable and a set input variables with deterministic connections equation. During simulation, theoretical approach used for initial assumptions about structure model boundaries measurement error intervals all being reliable not requiring verification their...

The class imbalance problem in classification scenarios is considered to be one of the main issues that limits performance many learning techniques. When reporting high accuracy a classifier may still exhibit poor for minority often interest. In this paper, we propose address by applying an SVM-based ensemble framework provides ability control trade-off between discovery rate under-represented classes and overall simultaneously. We evaluate proposed technique on both synthetic real-world...

The article is devoted to the problem of using remote sensing data for studying and mapping archaeological sites in interdisciplinary research. purpose experiments develop a methodology searching archeological monuments based on interpretation aerospace images. be solved formalized search procedure selecting objects. complex tasks ridentifying objects from images cannot realated only field decryption, it also deals with information processing signals (computer vision), this where great...

Геоинформационные и картографические методы технологии 199 for applied urban research], Moscow

The combination of multiple clustering solutions used to obtain accurate and novel output has attracted attention in data research. Despite the success ensembles, there are still several fundamental limiting issues including lack a unified formalized problem formulation an intuitive interpretation resulting solution. We formulate ensemble as binary matrix factorization imposing assumptions structure on matrices. In such framework, every object is assigned its representative centroid allowing...

Pattern matching in time series data streams is considered to be an essential mining problem that still stays challenging for many practical scenarios. Different factors such as noise, varying amplitude scale or shift, signal stretches shrinks are all leading performance degradation of existing pattern algorithms. In this paper, we introduce a dynamic z-normalization mechanism allowing proper scaling even under significant and distortions. Based on that, further propose Dynamic Time...

Within the framework of expanding technical and technological capabilities scientific methods, special interest are new ways obtaining field measurements, which can be successfully applied in combination with traditional archaeological survey. The technology unmanned shooting acts is one such methods. In course work on mapping sites establishment their boundaries, a comprehensive survey was developed tested. Intermediate final digital products have been created for all objects. first ones...

В данной статье описывается технология создания цифровой модели местности по материалам космической съемки. Исходными материалами для изготовления цифровых карт являлись космические снимки высокого разрешения формата GeoTiff, полученные со спутников QuickBird и Ikonos, с разрешением 0,6 1 м соответственно. результате описанной геоинформационной технологии были созданы цифровые карты М 1:5000 на 27 сел Усть-Коксинского района Республики Алтай. Ключевые слова: цифровая модель местности,...

The technology of geoinformational mapping and geoarchaeological analysis archaeological monuments is proposed, which makes it possible to analyze the distribution data on settlement population microdistrict Yustyt. A new program for calculating spatial correlation has been developed tested. Integral GIS models have built that reflect patterns

Searchable abstracts of presentations at key conferences in endocrinology ISSN 1470-3947 (print) | 1479-6848 (online)

Searchable abstracts of presentations at key conferences in endocrinology ISSN 1470-3947 (print) | 1479-6848 (online)

Searchable abstracts of presentations at key conferences in endocrinology ISSN 1470-3947 (print) | 1479-6848 (online)