• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.41 no.6
• /
• pp.19-24
• /
• 2004

A novel fast parallel NTGST is proposed for high resolution computer vision inspection of the BLUs in a LCD production line. The conventional computation- intensive NTGST algorithm is modified and its C codes are optimized into fast NTGST to be adapted to the SIMD parallel architecture. And then, the input inspection image is partitioned and allocated to each of the P processors in multi-threaded implementation, and the NTGST is executed on SIMD architecture of N data items simultaneously in each thread. Thus, the proposed inspection system can achieve the speedup of O(NP). Experiments using Dual-Pentium III processor with its MMX and extended MMX SIMD technology show that the proposed parallel NTGST is about Sp=8 times faster than the conventional NTGST, which shows the scalability of the proposed system implementation for the fast, high resolution computer vision inspection of the various sized BLUs in LCD production lines.

• The KIPS Transactions:PartC
• /
• v.9C no.4
• /
• pp.481-488
• /
• 2002

Guarantee of the data integrity is important to the Virtual Private Network (VPN) which can be improved cost decreasing and effective work by applying on Internet as the private network. Thus, the integrity function in the VPN must be maintained and the security manager must be check it occasionally. In this paper we propose the VPN Integrity Evaluation System (VIES) which is collecting, and evaluating automatically the vulnerable data of VPN against current hacking mechanisms in information security system. And this VIES obtain to the results which have objectivity and fairness of evaluation by driving off the evaluation scenario based on Common Criteria (CC), and general users or non-specialist can utilize easy the security evaluation of organization.

• KIPS Transactions on Software and Data Engineering
• /
• v.10 no.1
• /
• pp.9-18
• /
• 2021

Although Generative Adversarial Networks (GANs) have gained great popularity in computer vision and related fields, generating audio signals independently has yet to be presented. Unlike images, an audio signal is a sampled signal consisting of discrete samples, so it is not easy to learn the signals using CNN architectures, which is widely used in image generation tasks. In order to overcome this difficulty, GAN researchers proposed a strategy of applying time-frequency representations of audio to existing image-generating GANs. Following this strategy, we propose an improved method for increasing the fidelity of synthesized audio signals generated by using GANs. Our method is demonstrated on a public speech dataset, and evaluated by Fréchet Inception Distance (FID). When employing our method, the FID showed 10.504, but 11.973 as for the existing state of the art method (lower FID indicates better fidelity).

• KIPS Transactions on Software and Data Engineering
• /
• v.11 no.12
• /
• pp.509-516
• /
• 2022

Accurate pose prediction of objects in 3D space is an important visual recognition technique widely used in many applications such as scene understanding in both indoor and outdoor environments, robotic object manipulation, autonomous driving, and augmented reality. Most previous works for object pose estimation have the limitation that they require an exact 3D CAD model for each object. Unlike such previous works, this paper proposes a novel neural network model that can predict the poses of unknown objects based on only their RGB color images without the corresponding 3D CAD models. The proposed model can obtain depth maps required for unknown object pose prediction by using an adaptive depth estimator, AdaBins,. In this paper, we evaluate the usefulness and the performance of the proposed model through experiments using benchmark datasets.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.9 no.5
• /
• pp.213-224
• /
• 2009

This paper presents a new algorithm for estimating a driver's physical condition using real-time vision system and performs experimentation for real facial image data. The system relies on a face recognition to robustly track the center points and sizes of person's two pupils, and two side edge points of the mouth. The face recognition constitutes the color statistics by YUV color space together with geometrical model of a typical face. The system can classify the rotation in all viewing directions, to detect eye/mouth occlusion, eye blinking and eye closure, and to recover the three dimensional gaze of the eyes. These are utilized to determine the carelessness and drowsiness of the driver. Finally, experimental results have demonstrated the validity and the applicability of the proposed method for the estimation of a driver's physical condition.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.39A no.1
• /
• pp.1-11
• /
• 2014

In this paper we propose a system for vehicle detection and tracking which has the ability to learn on-line appearance changes of vehicles being tracked. The proposed system uses feature-based tracking method to estimate rapidly and robustly the motion of the newly detected vehicles between consecutive frames. Simultaneously, the system trains an online vehicle detector for the tracked vehicles. If the tracker fails, it is re-initialized by the detection of the online vehicle detector. An improved vehicle appearance model update rule is presented to increase a tracking performance and a speed of the proposed system. Performance of the proposed system is evaluated on the dataset acquired on various driving environment. In particular, the experimental results proved that the performance of the vehicle tracking is significantly improved under bad conditions such as entering a tunnel and passing rain.

• Journal of Broadcast Engineering
• /
• v.10 no.3
• /
• pp.286-296
• /
• 2005

This paper presents a stereo-based image compression and transmission system for 3D realistic television. In the proposed system, a disparity map is extracted from an input stereo image pair and the extracted disparity map and one of two input images are transmitted or stored at a local or remote site. However, correspondences can not be determined in occlusion areas. Thus, it is not easy to recover 3D information in such regions. In this paper, a reconstruction image compensation algorithm based on error block concealment and in-loop filtering is proposed to minimize the reconstruction error in generating stereo image pair. The effectiveness of the proposed algorithm is shown in term of objective accuracy of reconstruction image with several real stereo image pairs.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.24 no.1
• /
• pp.1-8
• /
• 2024

On-device AI technology, which can operate AI models at the edge devices to support real-time processing and privacy enhancement, is attracting attention. As intelligent IoT is applied to various industries, services utilizing the on-device AI technology are increasing significantly. However, general deep learning models require a lot of computational resources for inference and learning. Therefore, various lightweighting methods such as quantization and pruning have been suggested to operate deep learning models in embedded edge devices. Among the lightweighting methods, we analyze how to lightweight and apply deep learning models to edge computing devices, focusing on pruning technology in this paper. In particular, we utilize dynamic and static pruning techniques to evaluate the inference speed, accuracy, and memory usage of a lightweight AI vision model. The content analyzed in this paper can be used for intelligent video control systems or video security systems in autonomous vehicles, where real-time processing are highly required. In addition, it is expected that the content can be used more effectively in various IoT services and industries.

• Advanced Industrial SCIence
• /
• v.3 no.1
• /
• pp.23-29
• /
• 2024

In this scholarly investigation, the focus is placed on the transformative potential of edge computing in enhancing Intelligent Transportation Systems (ITS) for the facilitation of autonomous driving. The intrinsic capability of edge computing to process voluminous datasets locally and in a real-time manner is identified as paramount in meeting the exigent requirements of autonomous vehicles, encompassing expedited decision-making processes and the bolstering of safety protocols. This inquiry delves into the synergy between edge computing and extant ITS infrastructures, elucidating the manner in which localized data processing can substantially diminish latency, thereby augmenting the responsiveness of autonomous vehicles. Further, the study scrutinizes the deployment of edge servers, an array of sensors, and Vehicle-to-Everything (V2X) communication technologies, positing these elements as constituents of a robust framework designed to support instantaneous traffic management, collision avoidance mechanisms, and the dynamic optimization of vehicular routes. Moreover, this research addresses the principal challenges encountered in the incorporation of edge computing within ITS, including issues related to security, the integration of data, and the scalability of systems. It proffers insights into viable solutions and delineates directions for future scholarly inquiry.

• KIPS Transactions on Computer and Communication Systems
• /
• v.6 no.10
• /
• pp.407-414
• /
• 2017

Pig monitoring applications consisting of many tasks can take advantage of inherent data parallelism and enable parallel processing using performance accelerators. In this paper, we propose a task distribution method for pig monitoring applications into a heterogenous computing platform consisting of a multicore-CPU and a manycore-GPU. That is, a parallel program written in OpenCL is developed, and then the most suitable processor is determined based on the measured execution time of each task. The proposed method is simple but very effective, and can be applied to parallelize other applications consisting of many tasks on a heterogeneous computing platform consisting of a CPU and a GPU. Experimental results show that the performance of the proposed task distribution method on three different heterogeneous computing platforms can improve the performance of the typical GPU-only method where every tasks are executed on a deviceGPU by a factor of 1.5, 8.7 and 2.7, respectively.