• Smart Media Journal
• /
• v.13 no.6
• /
• pp.16-23
• /
• 2024

Pigs express their reactions to their environment and health status through a variety of sounds, such as grunting, coughing, and screaming. Given the significance of pig vocalizations, their study has recently become a vital source of data for livestock industry workers. To facilitate this, we propose a lightweight deep learning model based on MobileNet that analyzes pig vocal patterns to distinguish pig voices from farm noise and differentiate between vocal sounds and coughing. This model was able to accurately identify pig vocalizations amidst a variety of background noises and cough sounds within the pigsty. Test results demonstrated that this model achieved a high accuracy of 98.2%. Based on these results, future research is expected to address issues such as analyzing pig emotions and identifying stress levels.

• Journal of KIISE:Information Networking
• /
• v.33 no.2
• /
• pp.131-138
• /
• 2006

To achieve security in wireless sensor networks(WSN), it is important to be able to encrypt and authenticate messages sent among sensor nodes. Due to resource constraints, many key agreement schemes used in general networks such as Diffie-Hellman and public-key based schemes are not suitable for wireless sensor networks. The current pre-distribution of secret keys uses q-composite random key and it randomly allocates keys. But there exists high probability not to be public-key among sensor nodes and it is not efficient to find public-key because of the problem for time and energy consumption. To remove problems in pre-distribution of secret keys, we propose a new cryptographic key management protocol, which is based on the clustering scheme but does not depend on probabilistic key. The protocol can increase efficiency to manage keys because, before distributing keys in bootstrap, using public-key shared among nodes can remove processes to send or to receive key among sensors. Also, to find outcompromised nodes safely on network, it selves safety problem by applying a function of lightweight attack-detection mechanism.

• Journal of the Semiconductor & Display Technology
• /
• v.18 no.3
• /
• pp.122-126
• /
• 2019

These-days technology related to IoT (Internet of Thing) is widely used and there are many types of smart system based IoT like smart health, smart building and so on. In smart health system, it is possible to check someone's health by analyzing data from wearable IoT device like smart watch. Smart building system aims to collect data from sensor such as humidity, temperature, human counter like that and control the building for energy efficiency, security, safety and so forth. Furthermore, smart city system can comprise several smart systems like smart building, smart health, smart mobility, smart energy and etc. In this paper, we propose multimedia IoT based intelligent emergency alarm system for smart city. In existing IoT based smart system, it communicates lightweight data like text data. In the past, due to network's limitations lightweight IoT protocol was proposed for communicating data between things but now network technology develops, problem which is to communicate heavy data is solving. The proposed system obtains video from IP cameras/CCTVs, analyses the video by exploiting AI algorithm for detecting emergencies and prevents them which cause damage or death. If emergency is detected, the proposed system sends warning message that emergency may occur to people or agencies. We built prototype of the intelligent emergency alarm system based on MQTT and assured that the system detected dangerous situation and sent alarm messages. From the test results, it is expected that the system can prevent damages of people, nature and save human life from emergency.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.17 no.2
• /
• pp.93-102
• /
• 2007

A serious problem to fight DDoS attacks is that attackers use incorrect or spoofed IP addresses in the attack packets. Due to the stateless nature of the internet, it is a difficult problem to determine the source of these spoofed IP packets. The most of previous studies to prevent and correspond to DDoS attacks using the traceback mechanism have been accomplished in IPv4 environment. Even though a few studies in IPv6 environment were introduced, those have no detailed mechanism to cope with DDoS attacks. The mechanisms for tracing the origin of attacks in IPv6 networks have so many differences from those of IPv4 networks. In this paper we proposed a lightweight IP traceback mechanism in IPv6 network environment. When marking for traceback is needed, the router can generate Hop-by-Hop option and transmit the marked packet. We measured the performance of this mechanism and at the same time meeting the efficient marking for traceback.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.26 no.6
• /
• pp.813-826
• /
• 2022

AI-based image processing technologies in various fields have been widely studied. However, the lighter the board, the more difficult it is to reduce the weight of image processing algorithm due to a lot of computation. In this paper, we propose a method using deep learning for object recognition algorithm in lightweight embedded boards. We can determine the area using a deep neural network architecture algorithm that processes semantic segmentation with a relatively small amount of computation. After masking the area, by using more accurate deep learning algorithm we could operate object detection with improved accuracy for efficient neural network (ENet) and You Only Look Once (YOLO) toward executing object recognition in real time for lightweighted embedded boards. This research is expected to be used for autonomous driving applications, which have to be much lighter and cheaper than the existing approaches used for object recognition.

• Journal of Platform Technology
• /
• v.10 no.4
• /
• pp.39-46
• /
• 2022

Recently, digital contents including video and sound are created in various fields, transmitted to the cloud through the Internet, and then stored and used. In order to utilize digital content, it is essential to verify data integrity, and it is necessary to ensure network bandwidth efficiency of verified data. This paper describes the design and implementation of a server that maintains, manages, and provides data for verifying the integrity of video data. The server receives and stores image data from Logger, a module that acquires image data, and performs a function of providing data necessary for verification to Verifier, a module that verifies image data. Then, a lightweight Merkle tree is constructed using the hash value. The light-weight Merkle tree can quickly detect integrity violations without comparing individual hash values of the corresponding video frame changes of the video frame indexes of the two versions. A lightweight Merkle tree is constructed by generating a hash value of digital content so as to have network bandwidth efficiency, and the result of performing proof of integrity verification is presented.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2022.06a
• /
• pp.205-208
• /
• 2022

Multi-exposure high dynamic range (HDR) image reconstruction, the task of reconstructing an HDR image from multiple low dynamic range (LDR) images in a dynamic scene, often produces ghosting artifacts caused by camera motion and moving objects and also cannot deal with washed-out regions due to over or under-exposures. While there has been many deep-learning-based methods with motion estimation to alleviate these problems, they still have limitations for severely moving scenes. They also require large parameter counts, especially in the case of state-of-the-art methods that employ attention modules. To address these issues, we propose a frequency domain approach based on the idea that the transform domain coefficients inherently involve the global information from whole image pixels to cope with large motions. Specifically we adopt Residual Fast Fourier Transform (RFFT) blocks, which allows for global interactions of pixels. Moreover, we also employ Depthwise Overparametrized convolution (DO-conv) blocks, a convolution in which each input channel is convolved with its own 2D kernel, for faster convergence and performance gains. We call this LFFNet (Lightweight Frequency Fusion Network), and experiments on the benchmarks show reduced ghosting artifacts and improved performance up to 0.6dB tonemapped PSNR compared to recent state-of-the-art methods. Our architecture also requires fewer parameters and converges faster in training.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.12 no.4
• /
• pp.1832-1853
• /
• 2018

The personal health record (PHR) system is a promising application that provides precise information and customized services for health care. To flexibly protect sensitive data, attribute-based encryption has been widely applied for PHR access control. However, escrow, exposure and abuse of private keys still hinder its practical application in the PHR system. In this paper, we propose a coordinated ciphertext policy attribute-based access control with user accountability (CCP-ABAC-UA) for the PHR system. Its coordinated mechanism not only effectively prevents the escrow and exposure of private keys but also accurately detects whether key abuse is taking place and identifies the traitor. We claim that CCP-ABAC-UA is a user-side lightweight scheme. Especially for PHR receivers, no bilinear pairing computation is needed to access health records, so the practical mobile PHR system can be realized. By introducing a novel provably secure construction, we prove that it is secure against selectively chosen plaintext attacks. The analysis indicates that CCP-ABAC-UA achieves better performance in terms of security and user-side computational efficiency for a PHR system.

• ETRI Journal
• /
• v.32 no.5
• /
• pp.713-721
• /
• 2010

E-textile technology has earned a great deal of interest in many fields; however, existing wearable network protocols are not optimized for use with conductive yarn. In this paper, some of the basic properties of conductive textiles and requirements on wearable personal area networks (PANs) are reviewed. Then, we present a wearable personal network (WPN), which is a four-layered wearable PAN using bus topology. We have designed the WPN to be a lightweight protocol to work with a variety of microcontrollers. The profile layer is provided to make the application development process easy. The data link layer exchanges frames in a master-slave manner in either the reliable or best-effort mode. The lower part of the data link layer and the physical layer of WPN are made of a fabric serial-bus interface which is capable of measuring bus signal properties and adapting to medium variation. After a formal verification of operation and performances of WPN, we implemented WPN communication modules (WCMs) on small flexible printed circuit boards. In order to demonstrate the behavior of our WPN on a textile, we designed a WPN tutorial shirt prototype using implemented WCMs and conductive yarn.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2003.05a
• /
• pp.203-207
• /
• 2003

Mobile Ad Hoc Network is a kind of network technology, that mobile hosts can compose networks freely and temporarily where the base networks don't exist and where its establishment is difficult or as occasion demands. Ubiquitous computer network is prograssing in the forms of miniaturization, lightweight or portability. So far, MANET was studied in the point how to route it efficiently, but its power saving problems are been issuing nowadays. We try to not only compare and analyze power saving protocols in multi-hop Ad Hoc network but also present their efficiencies in this thesis.