Search

Along with the strong development of camera networks, a video analysis system has been become more and more popular and has been applied in various practical applications. In this paper, we focus on person re-identification (person ReID) task that is a crucial step of video analysis systems. The purpose of person ReID is to associate multiple images of a given person when moving in a non-overlapping camera network. Many efforts have been made to person ReID. However, most of studies on person ReID only deal with well-alignment bounding boxes which are detected manually and considered as the perfect inputs for person ReID. In fact, when building a fully automated person ReID system the quality of the two previous steps that are person detection and tracking may have a strong effect on the person ReID performance. The contribution of this paper are two-folds. First, a unified framework for person ReID based on deep learning models is proposed. In this framework, the coupling of a deep neural network for person detection and a deep-learning-based tracking method is used. Besides, features extracted from an improved ResNet architecture are proposed for person representation to achieve a higher ReID accuracy. Second, our self-built dataset is introduced and employed for evaluation of all three steps in the fully automated person ReID framework.

Along with the strong development of camera networks, a video analysis system has been become more and more popular and has been applied in various practical applications. In this paper, we focus on person re-identification (person ReID) task that is a crucial step of video analysis systems. The purpose of person ReID is to associate multiple images of a given person when moving in a non-overlapping camera network. Many efforts have been made to person ReID. However, most of studies on person ReID only deal with well-alignment bounding boxes which are detected manually and considered as the perfect inputs for person ReID. In fact, when building a fully automated person ReID system the quality of the two previous steps that are person detection and tracking may have a strong effect on the person ReID performance. The contribution of this paper are two-folds. First, a unified framework for person ReID based on deep learning models is proposed. In this framework, the coupling of a deep neural network for person detection and a deep-learning-based tracking method is used. Besides, features extracted from an improved ResNet architecture are proposed for person representation to achieve a higher ReID accuracy. Second, our self-built dataset is introduced and employed for evaluation of all three steps in the fully automated person ReID framework.

This study aims to investigate the influence of e-quality and online trust on customer engagement and e-word of mouth. In particular, this study explored and analyzed a relatively new relationship, the impact of customer engagement on e-word of mouth. The measurement model and conceptual model describing the relationships hypothesized in the study was evaluated, based on responses from 370 online purchasing customers who are students or office workers in Ho Chi Minh City. E-quality has a direct impact on online trust, which impacts online customer engagement of customers and e-word of mouth. Online trust has a direct effect on customer engagement and e-word-of-mouth. In particular, online engagement impacts on e-word of mouth. This study provides not only theoretical and practical meaning, and enables companies to realize the importance of customer engagement and e-word of mouth but also a number of solutions to help businesses build and increase their customer engagement and positive e-word of mouth.

Several parts of Binh Phuoc Province, southern Vietnam, suffer from degraded soils and vegetation as a result of both natural erosion of weak mud rocks and sandstones and intensive human activity, especially through land clearing for agriculture on unstable slopes, deforestation, and abandonment of poor farmland. The underlying cause of this land degradation has been the farming habits of migrants of varying ethnic groups who have settled in the area since 1980. The indigenous farming knowledge of these people and the role of that knowledge in soil erosion were examined by a series of household surveys. They enabled farming practices to be related to ground cover established from a 2002 Landsat 7 ETM (Enhanced Thematic Mapper), and erosion data from a series of erosion bridge measurements. A GIS (Geographical Information System) approach was piloted as a means of identifying areas vulnerable to erosion. This could then be combined with the understanding of farming practices to reveal the relative roles of farmer behaviour, crop cover, and slope and soil characteristics in the erosion process. Land use, local people's knowledge and economic realities are the main factors, as well as natural conditions, that drive this land degradation.

To help improve the well-being of the local people, a joint Vietnamese-UK team set out to establish a way of estimating soil and nutrient losses under different land management scenarios, using field data extrapolated through remote sensing and GIS, to obtain catchment-wide estimates of the impact of land cover change. Immigration from remote provinces to the Dong Phu District of Binh Phuóc Province, about 120 km north of Ho Chi Minh City, has led to disruption of soil surface stability on easily eroded clayey sandstones, creating rapid nutrient depletion that affects crop yields and siltation in the channel of the Rach Rat river downstream. The poor farmers of the areas see crop yields drop dramatically after two or three years of cultivation due to the fertility decline. Soil loss varies dramatically between wet season and dry season and with ground cover. Erosion bridge measurements showed a mean loss of 85.2 t ha-1 y-1 under cassava saplings with cashew nuts, 43.3 t ha-1 y-1 on uncultivated land and 41.7 t ha-1 y-1 under mature cassava. The rates of erosion were higher than those reported in many other parts of Vietnam, reflecting the high erodibility of the friable sandy soils on the steep side-slopes of the Rach Rat catchment. However, although the actual measurements provide better soil loss data than estimates based on the parameters of soil loss equations, a large number of measurement sites is needed to provide adequate coverage of the crop and slope combinations in this dissected terrain for good prediction using GIS and remote sensing. © 2008 The Authors Journal compilation © 2008 Institute of Australian Geographers.

1

1