• Journal of Information Processing Systems
• /
• v.15 no.4
• /
• pp.1017-1028
• /
• 2019

Pedestrian tracking is a particular object tracking problem and an important component in various vision-based applications, such as autonomous cars and surveillance systems. Following several years of development, pedestrian tracking in videos remains challenging, owing to the diversity of object appearances and surrounding environments. In this research, we proposed a tracking-by-detection system for pedestrian tracking, which incorporates a convolutional neural network (CNN) and color information. Pedestrians in video frames are localized using a CNN-based algorithm, and then detected pedestrians are assigned to their corresponding tracklets based on similarities between color distributions. The experimental results show that our system is able to overcome various difficulties to produce highly accurate tracking results.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.36 no.2
• /
• pp.259-268
• /
• 2016

The core strategy of transit oriented development (TOD) is to promote high density mixed land use around railway stations. Case studies in advanced countries show that provision of policies for comprehensive maintenance of pedestrian facilities around railway station spheres is being pursued with efficacy. In spite of the importance placed on integrated pedestrian maintenance, domestic construction of integrated pedestrian infrastructure around railway station spheres lacks direction. Thus, there is a clear need for an evaluation standard that can provide the foundation for judgments on TOD improvement. This research proposes a network model that consolidates the interior of the station as well as its surrounding areas to determine the ease of pedestrian flow for effective TOD evaluation. The model considers the railway station and surrounding areas as an assembled network of pedestrian flow. The path chosen by the pedestrian is defined as the optimal degree of inconvenience, and expands on Wardrop's User Equilibrium (1952). To assess the various circumstances that arise on pedestrian facilities including congestion of the pedestrian pathway, constrained elevator capacity, and wait at the crosswalk, a variational inequality based pedestrian equilibrium distribution model is introduced.

• International Journal of Internet, Broadcasting and Communication
• /
• v.8 no.1
• /
• pp.27-32
• /
• 2016

The global positioning system (GPS)-enabled mobile phones provide location-based applications such as car and pedestrian navigation services. The pedestrian navigation services provide safe and comfortable route and path guidance for pedestrians and handicapped or elderly people. One of the essential components for a navigation system is a spatial database used to perform navigation and routing functions. In this paper, we develop modeling and categorization of pedestrian path components for smart tour guide services using the mobile pedestrian navigation application. We create pedestrian networks using 2D base map and sky view map in urban area. We also construct pedestrian networks and attributes of node, link, and POI using on-site GPS data and photos for smart pedestrian tour guide in the major walking tourist spots in Jeju.

• Journal of the Architectural Institute of Korea Planning & Design
• /
• v.34 no.11
• /
• pp.85-94
• /
• 2018

This study aims to investigate how easy pedestrians get around within/through the "Apartment Complexes (AC), " a common style of high-rise multi-family housing in Korea. Over the past six decades, the AC has been the most conventional way to provide standardized housing efficiently to address the problems of the shortage of housing and the substandard housing, due to the explosion of urban population with the rapid industrialization. The AC is a huge chunk of homeogenous multi-family housing, mostly condos with decent infrastructure, including parks, pedestrian passages, schools, ect. Both in the new town development and urban renewal programs have utilized the advantages of the AC. Since the design principals of AC tend to adopt the "protective design" to prevent cars and pedestrians coming outside from passing it, it has been criticised for dissecting the continuity of socioeconomic context in neighborhoods. The neo-traditional planning urbanists, including Jane Jacobs, emphasize that smaller blocks and grid road newtworks are the key in improving social, cultural, and economic vitality of the neighborhoods, because these design concepts allow more pedestrians and different types of people to be mixed in a neighborhood. In this study, we first adopted objective measures for pedestrian accessibility and pedestrian efficiency. These measures were used to calculate the lengths of shortest paths from residential buildings to the edges of AC. We tested the difference in shortest paths between the current pedestrian networks of AC and hypothetical grid networks on the AC, and the relative difference is considered as the pedestrian efficiency, using the network analysis function of Geographic Information Systems (GIS) and Python programming. We found from the randomly selected 30 ACs that the existing non-grid road networks in ACs are worse than the hypothesized grid networks, in terms of pedestrian efficiency. In average, pedestrians in AC with the conventional road networks have to walk than 25%, 26%, and 27% longer than the networks of $125{\times}45m$, $100{\times}45m$, and $75{\times}45m$, respectively. With the t-test analysis, we found the pedestrian efficiency of AC with the conventional network is lower than grid-networks. Many new urbanists stress, easiness of walking is one of the most import elements for community building and social bonds. With the findings from the objective measures of pedestrian accessibility and efficiency, the AC would have limitations to attract people outside into the AC itself, which would increase dis-connectivity with adjacent areas.

• Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography
• /
• v.36 no.5
• /
• pp.335-341
• /
• 2018

The KDE (Kernel Density Estimation) technique in GIS (Geographic Information System) has been widely used as a method for determining whether a phenomenon occurring in space forms clusters. Most human-generated events such as traffic accidents and retail stores are distributed according to a road network. Even if events on forward and rear roads have short Euclidean distances, network distances may increase and the correlation between them may be low. Therefore, the NKDE (Network-based KDE) technique has been proposed and applied to the urban space where a road network has been developed. KDE is being studied in the field of business GIS, but there is a limit to the microscopic analysis of economic activity along a road. In this study, the NKDE technique is applied to the analysis of urban phenomena such as the density of shops rather than traffic accidents that occur on roads. The results of the NKDE technique are also compared to pedestrian networks and road centerline networks. The results show that applying NKDE to microscopic trade area analysis can yield relatively accurate results. In addition, it was found that pedestrian network data that can consider the movement of actual pedestrians are necessary for accurate trade area analysis using NKDE.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.12 no.1
• /
• pp.396-404
• /
• 2011

Recently, the wide spread popularity of smart phones causes the advent of various mobile applications base on the location information. Since previous pedestrian navigations are applied by extending car navigations, these are not only difficult to provide the appropriate route information, but also raise limitations in the efficient query processing by data structures of car road networks. In addition, these increase the power consumption caused by the growth of I/O frequency. In this paper, we propose a pedestrian road network model for the accurate route information and a storage structure for the pedestrian road network based on POI to reduce the I/O frequency. The proposed method enables efficient route searches over POI reflecting the characteristics and requirements of pedestrian roads. Also, a reduction of query processing costs for the route searching by a data structure considered with POI can save the power consumption more than previous approaches.

• The Journal of Korean Institute of Information Technology
• /
• v.16 no.12
• /
• pp.75-83
• /
• 2018

Recently, autonomous vehicles have been actively studied. Pedestrian detection and recognition technology is important in autonomous vehicles. Pedestrian detection using CNN(Convolutional Neural Netwrok), which is mainly used recently, generally shows good performance, but there is a performance degradation depending on the environment of the image. In this paper, we propose a pedestrian detection system applying long-term memory structure of hippocampal neural network based on CNN network with LGP-FL (Local Gradient Pattern-Feature Layer) added. First, change the input image to a size of $227{\times}227$. Then, the feature is extracted through a total of 5 layers of convolution layer. In the process, LGP-FL adds the LGP feature pattern and stores the high-frequency pattern in the long-term memory. In the detection process, it is possible to detect the pedestrian more accurately by detecting using the LGP feature pattern information robust to brightness and color change. A comparison of the existing methods and the proposed method confirmed the increase of detection rate of about 1~4%.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2016.10a
• /
• pp.506-509
• /
• 2016

We present real-time pedestrian detection that exploit accuracy of Faster R-CNN network. Faster R-CNN has shown to success at PASCAL VOC multi-object detection tasks, and their ability to operate on raw pixel input without the need to design special features is very engaging. Therefore, in this work we apply and adjust Faster R-CNN to single object detection, which is pedestrian detection. The drawback of Faster R-CNN is its failure when object size is small. Previously, small sized object problem was solved by Scale-aware Network. We incorporate Scale-aware Network to Faster R-CNN. This made our method Scale-aware Faster R-CNN (DF R-CNN) that is both fast and very accurate. We separated Faster R-CNN networks into two sub-network, that is one for large-size objects and another one for small-size objects. The resulting approach achieves a 28.3% average miss rate on the Caltech Pedestrian detection benchmark, which is competitive with the other best reported results.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.15 no.7
• /
• pp.2321-2338
• /
• 2021

Compared with vehicle trajectories, pedestrian trajectories have stronger degrees of freedom and complexity, which poses a higher challenge to trajectory prediction tasks. This paper designs a mode to divide the trajectory of pedestrians at a traffic intersection, which converts the trajectory regression problem into a trajectory classification problem. This paper builds a deep model for pedestrian trajectory prediction at intersections for the task of pedestrian short-term trajectory prediction. The model calculates the spatial correlation and temporal dependence of the trajectory. More importantly, it captures the interactive features among pedestrians through the Attention mechanism. In order to improve the training speed, the model is composed of pure convolutional networks. This design overcomes the single-step calculation mode of the traditional recurrent neural network. The experiment uses Vulnerable Road Users trajectory dataset for related modeling and evaluation work. Compared with the existing models of pedestrian trajectory prediction, the model proposed in this paper has advantages in terms of evaluation indicators, training speed and the number of model parameters.

• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.7 no.5
• /
• pp.26-32
• /
• 2008

A pedestrian navigation system, an information delivery server, and a program for naturally guiding (such as speech-guiding) the user of a portable terminal at an intersection. An information delivery server comprises a map database containing data such as nodes including paths constituting intersections, links, and costs of the links. The node-link structure is the most important part in pedestrian navigation system. Functional requirements for the road map database vary in different navigation phases. though there are various road network models, their traditional node-link structures, unfortunately, do not solve the problem well. This paper proposes a node-link structure for pedestrian navigation system. The network topological structure in pedestrianl network is presented, which accords with the practical walking habit better than traditional way treating the entire road network.