• Proceedings of the Korea Information Processing Society Conference
• /
• 2024.05a
• /
• pp.774-776
• /
• 2024

The effective implementation of advanced speech recognition (ASR) systems necessitates the deployment of sophisticated keyword spotting models that are both responsive and resource-efficient. The initial local detection of user interactions is crucial as it allows for the selective transmission of audio data to cloud services, thereby reducing operational costs and mitigating privacy risks associated with continuous data streaming. In this paper, we address these needs and propose utilizing the Whisper-Tiny model with fine-tuning process to specifically recognize keywords from google speech dataset which includes 65000 audio clips of keyword commands. By adapting the model's encoder and appending a lightweight classification head, we ensure that it operates within the limited resource constraints of local devices. The proposed model achieves the notable test accuracy of 92.94%. This architecture demonstrates the efficiency as on-device model with stringent resources leading to enhanced accessibility in everyday speech recognition applications.

• KIPS Transactions on Computer and Communication Systems
• /
• v.2 no.9
• /
• pp.381-392
• /
• 2013

During the last decade, the multitude of advances attained in terminal computers, along with the introduction of mobile hand-held devices, and the deployment of high speed networks have led to a recent surge of interest in Quality of Service (QoS) for video applications. The main difficulty is that mobile devices experience disparate channel conditions, which results in different rates and patterns of packet loss. One way of making more efficient use of network resources in video services over wireless channels with heterogeneous characteristics to heterogeneous types of mobile device is to use a scalable video coding (SVC). An SVC divides a video stream into a base layer and a single or multiple enhancement layers. We have to ensure that the base layer of the video stream is successfully received and decoded by the subscribers, because it provides the basis for the subsequent decoding of the enhancement layer(s). At the same time, a system should be designed so that the enhancement layer(s) can be successfully decoded by as many users as possible, so that the average QoS is as high as possible. To accommodate these characteristics, we propose an efficient transmission scheme which incorporates SVC-aware dual-channel repetition to improve the perceived quality of services. We repeat the base-layer data over two channels, with different characteristics, to exploit transmission diversity. On the other hand, those channels are utilized to increase the data rate of enhancement layer data. This arrangement reduces service disruption under poor channel conditions by protecting the data that is more important to video decoding. Simulations show that our scheme safeguards the important packets and improves perceived video quality at a mobile device.

• Asia pacific journal of information systems
• /
• v.20 no.2
• /
• pp.157-180
• /
• 2010

Rapid development of information technology has generated a new industry and market. In particular, network technology such as the Internet and other computer networks made transaction activities switched from traditional offline commerce to e-commerce. Among them, digital content is bit-based object which is created and distributed through electronic environment. In particular, many entertainment contents such as the music, the movies, and the computer game softwares are main products. Although digital content commerce has high potential demand, it lacks the consideration about the factor related to maintaining existing customer such as customer loyalty and switching costs. There has been a number of research on customer loyalty and other factors affecting it in the traditional electronic commerce environment, but there is a lack of research which examines the characteristics of digital content. The study about the effect of switching costs on customer loyalty in digital content commerce is necessary because the customers of digital content commerce market are from those of other e-commerce market or traditional offline commerce market. In addition, customer loyalty and switching costs are important factors because they may build up greater customer retention. For that reason, this study focused on examining the relationships among switching costs, antecedents for switching costs, and customer loyalty in online digital music service industry. The study has three major purposes: (1) to find antecedents of switching costs on digital content commerce and examine effect of antecedents for switching costs; (2) to identify effect of switching costs on customer loyalty in digital content commerce and examine moderating effects of alternative attractiveness; (3) to identify the differences of antecedents for switching costs by contents transmission type(streaming service and downloading service). And, the online digital music service industry is selected in this study since there are many users and transactions incurring. To accomplish these purposes, a survey questionnaire was developed and distributed to 256 informants. Survey instrument was developed based on previous research and pre-established survey items. Total of 206 surveys are collected and used in the data analysis. Among the respondents, 56.8% is male and 43.2% is female. Also, 86 responses were streaming service user group and 120 responses were download service user group. These data was analyzed using regression analysis. Major findings of empirical analysis can be summarized as follows. First, switching costs have positive effect on customer loyalty in digital content commerce environment. Second, the influence of switching costs on customer loyalty increases under conditions of high alternative attractiveness. Third, DRM convenience and breadth of use have positive effect on switching costs. The findings imply that the digital content provider should pay more attention to switching costs in addition to customer satisfaction in order to attract customers. Also, increasing the convenience of DRM use by securing the convenience of user interface and expanding the support device and increasing the service use scope by providing diverse value-added service helps to create a switching barrier. The result of the study can become a practical use in marketing strategy for maintaining existing customer. In particular, switching barrier is very important under conditions of high competition in the online music service market. This study can be used as a basis for further studies about customer retention in digital content commerce.

• 대한공업교육학회지
• /
• v.33 no.2
• /
• pp.287-299
• /
• 2008

HTTP is one of the most widely used protocols of the WWW. Currently it uses TCP as the transport layer protocol to provide reliability. The HTTP uses separate TCP connection for each file request and adds unnecessary head-of-line blocking overhead for the file retrieval. The web application is short sized and affected by the increased handover latency of TCP in wireless environment. SCTP has attractive features such as multi-streaming and multi-homing. SCTP's multi-streaming and multi-homing avoid head-of-line blocking problem of TCP and reduce handover latency of TCP in wired and wireless environment. Mean response time is the important measure in most web application. In this paper, we present the comparison of mean response time between HTTP over SCTP with that of HTTP over TCP in wired and wireless environments using NS-2 simulator. We measured mean response time for varying packet loss rate, bandwidth, RTT, and the number of web objects in wired environment and mean response time and packet loss rate for varying moving speed and region size in wireless environment. Our experimental result shows that SCTP reduces the mean response time of TCP based web traffic.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.30 no.11A
• /
• pp.1081-1091
• /
• 2005

Unlike conventional DVB-T transmission with the streaming method, DVB-H system based on the IPDC(IP Data Casting) method uses Time-slicing scheme to achieve the maximization of portability by reducing the power consumption of a receiver. To enhance the power efficiency of the receiver, Time-slicing scheme controls the receiver operation to perform only for corresponding burst in specific time slot. The additional power saving can also be achieved by reducing the required time for synchronization. In this paper, we propose a fast scattered pilot synchronization algorithm, which detects the pilot pattern of currently received OFDM symbol. The proposed scheme is based on the correlation between the adjacent subcarriers of potential scattered pilot position in two consecutively received OFDM symbols. Therefore, it can reduce the time for the scattered pilot synchronization within two symbols as com-pared with the conventional method used for DVB-T. And the proposed algorithm has better performance than the two schemes proposed by Nokia for DVB-H and the method using correlation with reference signal. Extensive com-puter simulation is performed based on ETSI EN300 744 ETSI and performance results show that the proposed algorithm has more efficient and stable operation than the conventional schemes.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2015.10a
• /
• pp.205-208
• /
• 2015

Large vessels are currently operating by using an ECDIS, a system enabling automatic navigation and checking marine conditions. In the past, the ECDIS was used for vessels but in these days, it has been used as a method for collecting and verifying marine information, etc. from land by using a computer, instead of taking a boat out. The ECDIS on sale is generally used by companies which operate marine workers. However, this system has caused inconvenience to marine workers, in terms of tasks, services, or operations. Marine workers move out to the sea after checking the status of the system at the office. Due to the weather such as waves and winds, marine devices are not often found on the relevant coordinates. In such a case, the workers onerously need to come back to the land and track the marine devices again. In order to solve such a problem, this thesis purposes to develop a system which can check the locations of marine devices and verify a marine chart even on the sea by using a mobile. The proposed system can integrate and support a variety of protocols of relevant devices in the system itself and check data via mobile through the data transmission and streaming functions.

• Journal of Broadcast Engineering
• /
• v.24 no.5
• /
• pp.849-858
• /
• 2019

Recent developments in computer graphics technology and image processing technology have increased interest in point cloud technology for inputting real space and object information as three-dimensional data. In particular, point cloud technology can accurately provide spatial information, and has attracted a great deal of interest in the field of autonomous vehicles and AR (Augmented Reality)/VR (Virtual Reality). However, in order to provide users with 3D point cloud contents that require more data than conventional 2D images, various technology developments are required. In order to solve these problems, an international standardization organization, MPEG(Moving Picture Experts Group), is in the process of discussing efficient compression and transmission schemes. In this paper, we provide a region division transfer method of 3D point cloud content through extension of existing MPEG-DASH (Dynamic Adaptive Streaming over HTTP)-SRD (Spatial Relationship Description) technology, quality parameters are further defined in the signaling message so that the quality parameters can be selectively determined according to the user's request. We also design a verification platform for ROUTE (Real Time Object Delivery Over Unidirectional Transport)/DASH based heterogeneous network environment and use the results to validate the proposed technology.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2017.04a
• /
• pp.661-664
• /
• 2017

Recently, during disasters occurrence, dealing with emergencies has been handled well by the early transmission of disaster relating notifications on social media networks (e.g., Twitter or Facebook). Intuitively, with their characteristics (e.g., real-time, mobility) and big communities whose users could be regarded as volunteers, social networks are proved to be a crucial role for disasters response. However, the amount of data transmitted during disasters is an obstacle for filtering informative messages; because the messages are diversity, large and very noise. This large volume of data could be seen as Social Big Data (SBD). In this paper, we proposed a big data platform for collecting and analyzing disasters' data from SBD. Firstly, we designed a collecting module; which could rapidly extract disasters' information from the Twitter; by big data frameworks supporting streaming data on distributed system; such as Kafka and Spark. Secondly, we developed an analyzing module which learned from SBD to distinguish the useful information from the irrelevant one. Finally, we also designed a real-time visualization on the web interface for displaying the results of analysis phase. To show the viability of our platform, we conducted experiments of the collecting and analyzing phases in 10 days for both real-time and historical tweets, which were about disasters happened in South Korea. The results prove that our big data platform could be applied to disaster information based systems, by providing a huge relevant data; which can be used for inferring affected regions and victims in disaster situations, from 21.000 collected tweets.

• Journal of Korea Multimedia Society
• /
• v.12 no.1
• /
• pp.69-84
• /
• 2009

In this paper, a real-time web-casting framework for many individuals or small organizations imploring internet broadcasting, is designed and implemented. It is the purpose of the proposed framework that any individual can easily construct his own broadcast station if he has minimum facility for the content transmission to the framework. This type of system should have a flexible architecture to be applicable in wide variation of operational environment. Since the proposed framework follows a component-based hierarchical structure consisting of 3 types of components, it can effectively cope with change of internal or external operational environment by simple modification of the associated component with the change. Also, various types of target application systems in the aspects of cost and performance can be constructed in accordance with how to associate upper layer components with hardware platforms. A real prototype system based on the proposed framework is implemented, and then the response delay time is measured under varying number of content productions and content receivers. The results show that the response delay time follows M/M/1 model, which is one of the well-known performance models, and the implemented system may accommodate from 16 to 20 productions.

• The KIPS Transactions:PartC
• /
• v.13C no.6 s.109
• /
• pp.767-774
• /
• 2006

In order to support overlay multicast with scalability and reliability, in this paper, we introduced a new multicast tree construction method and a regional multicast. The architecture we introduce efficiently transmits streaming data to the users from multicast source in an environment in which non-multicast network and multicast network. The new multicast tree construction method enables the connection with a remote relay which is an optimum condition in case there is no relay in subnet. Besides, by defining a new service type to traffic class of IPv6 header, a selective adoption or FEC and a distinctive transmission became possible. In order to verify the effect of the solution we suggested, we embodied an environment which is similar to the actual service environment by a simulator and tested it. The result of the test shows that the proposed method is more efficient than the traditional unicast approach method and the existing overlay multicast.