• Proceedings of the Korea Information Processing Society Conference
• /
• 2009.11a
• /
• pp.375-376
• /
• 2009

모바일 환경 상에서 영상통화는 제한된 대역폭 및 손떨림 현상 등의 문제로 인하여 저품질의 서비스가 제공되고 있다. 기존의 영상통화 환경에서 향상된 품질의 영상통화 서비스를 제공하기 위하여 중요 객체와 비중요 객체를 구분하여 분리하며, 중요 객체에 대해서는 고화질을 갖도록 하는 비디오 압축 방법을 제안한다. 매 프레임 마다 중요 객체와 비중요 객체를 분리하며, 분리된 객체는 다른 비트율과 화질을 갖도록 각 객체에 적합한 변환 커널을 적용하여 부호화 절차를 달리한다. 중요 객체는 저압축, 고화질을 가지도록 부호화하며, 비중요 객체에 대해서는 고압축, 저화질을 가지도록 부호화함으로써 제한된 대역폭 내에서 비트율을 유지하도록 한다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2022.05a
• /
• pp.576-578
• /
• 2022

딥러닝 기술의 발전으로 객체 인색, 영상 분석에 관한 성능이 비약적으로 발전하였다. 하지만 고성능 GPU 를 사용하는 컴퓨팅 환경이 아닌 제한적인 엣지 디바이스 환경에서의 영상 처리 및 딥러닝 모델의 적용을 위해서는 엣지 디바이스에서 딥러닝 모델 실행 환경 과 이에 대한 분석이 필요하다. 본 논문에서는 RISC-V ISA 를 구현한 RISC-V 가상 플랫폼에 yolov3-tiny 모델 기반 객체 인식 시스템을 소프트웨어 레벨에서 포팅하여 구현하고, 샘플 이미지에 대한 네트워크 딥러닝 연산 및 객체 인식 알고리즘을 적용하여 그 결과를 도출하여 보았다. 본 적용을 바탕으로 RISC-V 기반 임베디드 엣지 디바이스 플랫폼에서 딥러닝 네트워크 연산과 객체 인식 알고리즘의 수행에 대한 분석과 딥러닝 연산 최적화를 위한 알고리즘 연구에 활용할 수 있다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2023.05a
• /
• pp.79-80
• /
• 2023

CPU 와 GPU 의 성능이 지속적으로 발전함에 따라 객체 인식 인공지능의 정확도와 추론 속도는 점차 향상되고 있으나 이러한 성능을 Edge Device 와 같은 제한된 환경에서 구현하기에 아직 여러 한계점이 존재한다. 본 논문에서는 여러가지 Edge Device 에서 객체 인식을 위한 경량화 된 모델 중 하나인 SSD-Mobilenet-V2 를 활용하여 결과값을 통해 각 Device 간 경향성을 분석하였다. 본 결과를 바탕으로 다양한 환경에서의 객체인식 인공지능 모델의 성능 개선을 위한 연구에 활용할 수 있다.

• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.43 no.3 s.309
• /
• pp.20-29
• /
• 2006

The Java language has been established as one of the most widely used language due to its object-oriented programming, safety and flexibility and the garbage collection of the virtual machine has relieved programmers of many difficulties related to the memory management. In the embedded environment, Java is also prevalent, the virtual machine and garbage collector that takes into account the embedded environment is required. In this paper we manage the heap memory area by dividing into young generation and old generation, and we propose a garbage collector in which appropriate techniques are applied to each generation to utilize the different characteristics of each generation. Also, we propose the write barrier technique and double filtering technique for efficient garbage recognition, and double check method for determining and reclaiming the garbage with cyclic structure. The proposed method satisfies the embedded environment's requirements of fast object allocation, real time property, recollection of all the garbage, elimination of fragmentations and high locality.

• Korean Journal of Optics and Photonics
• /
• v.18 no.6
• /
• pp.410-420
• /
• 2007

We theoretically derive the set of utilizable paraxial zoom locus equations for all complicated zoom lens systems with infinite object distance, such as a camera zoom lens, by using the Gaussian bracket method and the matrix representation of paraxial ray tracing. And we make the zoom locus program according to these equations in Visual Basic. Since we have applied the paraxial ray tracing equations into Gaussian bracket representation, the resultant program systematically simplifies various constraints of the zoom loci of various N group types. Consequently, the solutions of this method can be consistently used in all types of zoom lens in the step of initial design about zoom loci. Finally, in order to verify the usefulness of this method, we show that one example among 4 groups and that among 5 groups, which are very complex zoom lens systems, can be rapidly and with versatility traced through various interpolations by using this program.

• Journal of Radiation Industry
• /
• v.5 no.4
• /
• pp.325-330
• /
• 2011

To predict the details of flow in industrial process unit, single photon emission computed tomography (SPECT) is a promising technique. Recently, industrial SPECT based on medical system has developed by researchers of the Korea Atomic Energy Research Institute (KAERI) and Hanyang University. In the present study, to confirm the effect of the number of detectors on image quality, and determine the optimal number of detectors in industrial SPECT, industrial SPECT system with various geometries were evaluated by the Monte Carlo simulation. CsI(Tl) detectors ($12mm{\times}12mm{\times}20mm$) with collimators (the geometric resolution of collimator $R_g$ was 4 cm at the center of the 30 cm diameter cylindrical vessel object) were modeled in a hexagonal array, and the point sources of $^{99m}Tc$, $^{68}Ga$, and $^{137}Cs$ were simulated at the center of the cylindrical vessel object using the MCNPX code. Then, the reconstruction images of each geometry were reconstructed using the expectation maximization (EM) algorithm. In this study, the reciprocity theorem was used to improve computation time required for system matrix of the EM algorithm. The result shows that the resolution of the reconstructed image was significantly improved by increasing the number of detectors in industrial SPECT system and more than 60 detectors will be required for the resolution of the reconstructed image.

• Journal of the Society of Naval Architects of Korea
• /
• v.60 no.4
• /
• pp.248-255
• /
• 2023

In order to implement the autonomous navigation function, it is essential to track an object within a certain radius of the ship's route. This paper proposes the Multiple Probabilistic Data Association Filter (MPDAF), which can track multiple ships by extending Probabilistic Data Association Filter (PDAF), an existing single object tracking algorithm, using radar data obtained from real marine environments. The proposed MPDAF algorithm was developed to address the problem of tracking multiple objects in a complex environment where there can be significant uncertainty in the number and identification of objects to be tracked. Using real-world radar data provided by the German aerospace center (DLR), it has been verified that the proposed algorithm can track a large number of objects with a small position error.

• Journal of information and communication convergence engineering
• /
• v.1 no.3
• /
• pp.99-103
• /
• 2003

Classes are apt to include useless codes and inadequate inheritance relationship between them when they are being updated, inserted and deleted during the evolution process of object-oriented software, leading to lots of errors. Conventional class restructuring methods degrade the effectiveness of reusability since they go with preprocesses such as dependency analysis and estimation of class cohesion and run statically. In this paper, we propose a new C＋＋ class-restructuring algorithm that does not require those preprocesses and runs dynamically by improving ART learning algorithm in the artificial neural networks.

• ETRI Journal
• /
• v.30 no.1
• /
• pp.152-154
• /
• 2008

We present a simple and efficient technique for a plausible elastic deformation of three-dimensional objects. An elastic sweep surface is constructed by interpolating key cross sections with positions, orientations, and boundary shapes determined by physical simulation of simple mass-spring systems. The deformable parts of an object are approximated by the elastic sweep surfaces, and the vertices of the deformable parts are bound to nearby sweep surfaces. As an external force is applied, the corresponding parts of an object change their shapes elastically. We demonstrate the effectiveness of our technique and show its real-time performance on mesh objects.

• International Journal of Control, Automation, and Systems
• /
• v.4 no.2
• /
• pp.217-226
• /
• 2006

Grasping and manipulation by hands can be considered as one of inevitable functions to achieve the performances desired in humanoid operations. When a humanoid robot manipulates an object by his hands, each finger should be well-controlled to accomplish a precise manipulation of the object grasped. So, the trajectory of each joint required for a precise finger motion is fundamentally necessary to be planned stably. In this sense, this paper proposes an effective joint motion planning method for humanoid fingers. The proposed method newly employs a bio-mimetic concept for joint motion planning. A suitable model that describes an interphalangeal coordination in a human finger is suggested and incorporated into the proposed joint motion planning method. The feature of the proposed method is illustrated by simulation results. As a result, the proposed method is useful for a facilitative finger motion. It can be applied to improve the control performance of humanoid fingers or prosthetic fingers.