• KIPS Transactions on Software and Data Engineering
• /
• v.12 no.12
• /
• pp.505-518
• /
• 2023

3D point cloud semantic segmentation is a computer vision task that involves dividing the point cloud into different objects and regions by predicting the class label of each point. Existing 3D semantic segmentation models have some limitations in performing sufficient fusion of multi-modal features while ensuring both characteristics of 2D visual features extracted from RGB images and 3D geometric features extracted from point cloud. Therefore, in this paper, we propose MMCA-Net, a novel 3D semantic segmentation model using 2D-3D multi-modal features. The proposed model effectively fuses two heterogeneous 2D visual features and 3D geometric features by using an intermediate fusion strategy and a multi-modal cross attention-based fusion operation. Also, the proposed model extracts context-rich 3D geometric features from input point cloud consisting of irregularly distributed points by adopting PTv2 as 3D geometric encoder. In this paper, we conducted both quantitative and qualitative experiments with the benchmark dataset, ScanNetv2 in order to analyze the performance of the proposed model. In terms of the metric mIoU, the proposed model showed a 9.2% performance improvement over the PTv2 model using only 3D geometric features, and a 12.12% performance improvement over the MVPNet model using 2D-3D multi-modal features. As a result, we proved the effectiveness and usefulness of the proposed model.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.14 no.4
• /
• pp.671-678
• /
• 2019

The semantic information provided through the semantic-based IoT system will produce new high value-added products that are completely different from what we have known and experienced. From this point of view, the key issue of current IoT technology and applications is the development of an intelligent IoT platform architecture. The proposed system collects the IoT data of the sensors from the cloud computer, converts them into RDF, and annotates them with semantics. The converted semantic data is shared and utilized through the ontology repository. We use KT's IoTMakers as a cloud computing environment, and the ontology repository uses Jena's Fuseki server to express SPARQL query results on the web using Daum Map API and Highcharts API. This gives people the opportunity to access the semantic IoT mash-up service easily and has various application possibilities.

• Korean Journal of Remote Sensing
• /
• v.39 no.2
• /
• pp.143-156
• /
• 2023

To precisely and reliably analyze the contents of the satellite imagery, recognizing the clouds which are the obstacle to gathering the useful information is essential. In recent times, deep learning yielded satisfactory results in various tasks, so many studies using deep neural networks have been conducted to improve the performance of cloud detection. However, existing methods for cloud detection have the limitation on increasing the performance due to the adopting the network models for semantic image segmentation without modification. To tackle this problem, we introduced the multi-branch neural architecture search to find optimal network structure for cloud detection. Additionally, the proposed method adopts the soft intersection over union (IoU) as loss function to mitigate the disagreement between the loss function and the evaluation metric and uses the various data augmentation methods. The experiments are conducted using the cloud detection dataset acquired by Arirang-3/3A satellite imagery. The experimental results showed that the proposed network which are searched network architecture using cloud dataset is 4% higher than the existing network model which are searched network structure using urban street scenes with regard to the IoU. Also, the experimental results showed that the soft IoU exhibits the best performance on cloud detection among the various loss functions. When comparing the proposed method with the state-of-the-art (SOTA) models in the field of semantic segmentation, the proposed method showed better performance than the SOTA models with regard to the mean IoU and overall accuracy.

• Korean Journal of Remote Sensing
• /
• v.38 no.5_3
• /
• pp.939-951
• /
• 2022

The most important factor in designing autonomous driving systems is to recognize the exact location of the vehicle within the surrounding environment. To date, various sensors and navigation systems have been used for autonomous driving systems; however, all have limitations. Therefore, the need for high-definition (HD) maps that provide high-precision infrastructure information for safe and convenient autonomous driving is increasing. HD maps are drawn using three-dimensional point cloud data acquired through a mobile mapping system (MMS). However, this process requires manual work due to the large numbers of points and drawing layers, increasing the cost and effort associated with HD mapping. The objective of this study was to improve the efficiency of HD mapping by segmenting semantic information in an MMS point cloud into six classes: roads, curbs, sidewalks, medians, lanes, and other elements. Segmentation was performed using various machine learning techniques including random forest (RF), support vector machine (SVM), k-nearest neighbor (KNN), and gradient-boosting machine (GBM), and 11 variables including geometry, color, intensity, and other road design features. MMS point cloud data for a 130-m section of a five-lane road near Minam Station in Busan, were used to evaluate the segmentation models; the average F1 scores of the models were 95.43% for RF, 92.1% for SVM, 91.05% for GBM, and 82.63% for KNN. The RF model showed the best segmentation performance, with F1 scores of 99.3%, 95.5%, 94.5%, 93.5%, and 90.1% for roads, sidewalks, curbs, medians, and lanes, respectively. The variable importance results of the RF model showed high mean decrease accuracy and mean decrease gini for XY dist. and Z dist. variables related to road design, respectively. Thus, variables related to road design contributed significantly to the segmentation of semantic information. The results of this study demonstrate the applicability of segmentation of MMS point cloud data based on machine learning, and will help to reduce the cost and effort associated with HD mapping.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.10 no.1
• /
• pp.38-57
• /
• 2016

Service discovery is one of the major challenges in cloud computing environment with a large number of service providers and heterogeneous services. Non-uniform naming conventions, varied types and features of services make cloud service discovery a grueling problem. With the proliferation of cloud services, it has been laborious to find services, especially from Internet-based service repositories. To address this issue, services are crawled and clustered according to their similarity. The clustered services are maintained as a catalogue in which the data published on the cloud provider's website are stored in a standard format. As there is no standard specification and a description language for cloud services, new efficient and intelligent mechanisms to discover cloud services are strongly required and desired. This paper also proposes a key-value representation to describe cloud services in a formal way and to facilitate matching between offered services and demand. Since naïve users prefer to have a query in natural language, semantic approaches are used to close the gap between the ambiguous user requirements and the service specifications. Experimental evaluation measured in terms of precision and recall of retrieved services shows that the proposed approach outperforms existing methods.

• Journal of the Korea Society of Computer and Information
• /
• v.22 no.6
• /
• pp.57-68
• /
• 2017

The Linked Open Data(LOD) cloud is quickly becoming one of the largest collections of interlinked datasets and the de facto standard for publishing, sharing and connecting pieces of data on the Web. Data publishers from diverse domains publish their data using Resource Description Framework(RDF) data model and provide SPARQL endpoints to enable querying their data, which enables creating a global, distributed and interconnected dataspace on the LOD cloud. Although it is possible to extract structured data as query results by using SPARQL, users have very poor in analysis and visualization of RDF data from SPARQL query results. Therefore, to tackle this issue, based on Formal Concept Analysis, we propose a novel approach for analyzing and visualizing useful information from the LOD cloud. The RDF data analysis and visualization technique proposed in this paper can be utilized in the field of semantic web data mining by extracting and analyzing the information and knowledge inherent in LOD and supporting classification and visualization.

• International Journal of Advanced Culture Technology
• /
• v.8 no.3
• /
• pp.223-230
• /
• 2020

Companies have made an investment of cost and time to optimize processing of a new business model in a cloud environment, applying collaboration technology utilizing business processes in a social network. The collaborative processing method changed from traditional BPM to the cloud and a mobile cloud environment. We proposed a collaborative system for operating processes in social networks using MOXMDR-DAI+ (eXtended Metadata Registry-Data Access & Integration based multimedia ontology). The system operating cloud-based collaborative processes in application of MOXMDR-DAI+, which was suitable for data interoperation. MOXMDR-DAI+ applied to this system was an agent effectively supporting access and integration between multimedia content metadata schema and instance, which were necessary for data interoperation, of individual local system in the cloud environment, operating collaborative processes in the social network. In operating the social network-based collaborative processes, there occurred heterogeneousness such as schema structure and semantic collision due to queries in the processes and unit conversion between instances. It aimed to solve the occurrence of heterogeneousness in the process of metadata mapping using MOXMDR-DAI+ in the system. The system proposed in this study can visualize business processes. And it makes it easier to operate the collaboration process through mobile support. Real-time status monitoring of the operation process is possible through the dashboard, and it is possible to perform a collaborative process through expert search using a community in a social network environment.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.17 no.8
• /
• pp.1996-2015
• /
• 2023

High-definition (HD) maps can provide precise road information that enables an autonomous driving system to effectively navigate a vehicle. Recent research has focused on leveraging semantic segmentation to achieve automatic annotation of HD maps. However, the existing methods suffer from low recognition accuracy in automatic driving scenarios, leading to inefficient annotation processes. In this paper, we propose a novel semantic segmentation method for automatic HD map annotation. Our approach introduces a new encoder, known as the convolutional transformer hybrid encoder, to enhance the model's feature extraction capabilities. Additionally, we propose a multi-level fusion module that enables the model to aggregate different levels of detail and semantic information. Furthermore, we present a novel decoupled boundary joint decoder to improve the model's ability to handle the boundary between categories. To evaluate our method, we conducted experiments using the Bird's Eye View point cloud images dataset and Cityscapes dataset. Comparative analysis against stateof-the-art methods demonstrates that our model achieves the highest performance. Specifically, our model achieves an mIoU of 56.26%, surpassing the results of SegFormer with an mIoU of 1.47%. This innovative promises to significantly enhance the efficiency of HD map automatic annotation.

• Journal of the Korea Society of Computer and Information
• /
• v.21 no.5
• /
• pp.11-18
• /
• 2016

Linked Data is a web standard data definition method devised to connect, expand resources with a standardized type. Linked Data built in various areas expands existing knowledge through an open data cloud like LOD(Linked Open Data). A project to link and service existing knowledge through LOD is under way worldwide. However, LOD project in domestic is being participated in a specific field to the level of research. In this paper, we suggests a method to build the area of technical knowledge like legislations in type of Linked Data, and distribute such Linked Data built to LOD. The construction method suggested by this paper divides knowledge of legislations in structural, semantic, and integrated perspective, and builds each of them by converting to Linked Data according to the perspective. Also, such built Linked Legal Data prepares to link knowledge in a standardized type by distributing them onto LOD. Built Linked Legal Data are equipped with schema for link service in various types, and give help increase understand the access type to existing legal information.

• KIPS Transactions on Software and Data Engineering
• /
• v.1 no.1
• /
• pp.19-30
• /
• 2012

The context aware service is the service to provide useful information to the users by recognizing surroundings around people who receive the service via computer based on computing and communication, and by conducting self-decision. But CAS(Context Awareness System) shows the weak point of small-scale context awareness processing capacity due to restricted mobile function under the current mobile environment, memory space, and inference cost increment. In this paper, we propose a mobile cloud context system with using Google App Engine based on PaaS(Platform as a Service) in order to get context service in various mobile devices without any subordination to any specific platform. Inference design method of the proposed system makes use of knowledge-based framework with semantic inference that is presented by SWRL rule and OWL ontology and Jess with rule-based inference engine. As well as, it is intended to shorten the context service reasoning time with mapping the regular reasoning of SWRL to Jess reasoning engine by connecting the values such as Class, Property and Individual which are regular information in the form of SWRL to Jess reasoning engine via JessTab plug-in in order to overcome the demerit of queries reasoning method of SparQL in semantic search which is a previous reasoning method.