• Information and Communications Magazine
• /
• v.15 no.12
• /
• pp.50-64
• /
• 1998

• Journal of Korea Multimedia Society
• /
• v.13 no.4
• /
• pp.560-574
• /
• 2010

Recently, many people can easily contact video telephony service through the mobile terminal owing to the commercialization of 3G communication technology. However, the quality of the serviced video telephony has been not good yet by the actual mobile restrictions. For solving quality problems, this paper presents the design of the object-based video CODEC using hybrid transform in mobile video telephony. The proposed design firstly segment each frame into a significant object and an insignificant object. The proposed design is to improve the quality of a significant object by limiting the bit rate of a insignificant object. Thus, we compress a significant object with high quality and low compression ratio and compress an insignificant object with low quality and high compression ratio. Furthermore, we control the bit rate of the video stream in the limited bandwidth by adjusting the compression ratio of each object. From experimental results, we confirmed that our method has more higher quality than methods in the conventional CODECs at the significant region on the same bit rate.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.28 no.2C
• /
• pp.147-161
• /
• 2003

Video object segmentation is one of the core technologies for content-based real-time MPEG-4 encoding system. For real-time requirement, the segmentation algorithm should be fast and accurate but almost all existing algorithms are computationally intensive and not suitable for real-time applications. The MPEG-4 VM(Verification Model) has provided basic algorithms for MPEG-4 encoding but it has many limitations in practical software development, real-time camera input system and compression efficiency. In this paper, we implemented the preprocessing system for real-time camera input and VOP extraction for content-based video coding and also implemented motion detection to achieve the 180 : 1 compression rate for real-time and high compression MPEG-4 encoding.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2010.11a
• /
• pp.175-177
• /
• 2010

본 논문에서는 MPEG 의 3차원 비디오 시스템의 표준 깊이정보 맵에 대한 효율적인 부호화를 위하여 전처리 방법을 제안한다. 현재 3차원 비디오 부호화(3DVC)에 대한 표준화가 진행 중에 있지만 아직 깊이정보 맵의 부호화 방법에 대한 표준이 확정되지 않은 상태이다. 제안하는 기법에서는 우선, 입력된 깊이정보 맵에 대하여 원래의 히스토그램 분포를 가우시안 혼합모델(GMM)기반의 EM 군집화 기법에 의한 방법으로 분리 후, 분리된 히스토그램을 기반으로 깊이정보 맵을 여러 개의 영상으로 분리한다. 그 후 분리된 각각의 영상을 배경과 객체에 따라 다른 조건의 mean shift filter로 필터링한다. 결과적으로 영상내의 각 영역 경계는 최대한 살리면서 영역내의 화소 값에 대해서는 평균 연산을 취하여 부호화시 효율을 극대화 하고자 하였다. 실험조건은 $1024{\times}768$ 영상에 대해서 50 프레임으로 H.264/AVC base 프로파일로 부호화를 진행하였다. 최종 실험결과 bit rate는 대략 23% ~ 26% 정도 감소하고 부호화 시간도 다소 줄어드는 것을 확인 할 수 있었다.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.41 no.1
• /
• pp.47-56
• /
• 2004

This paper describes a new bit allocation algorithm that can achieve a constant bit rate when coding multiple video objects, while improving rate-distortion (R-D) performance over the VM5 method for MPEG-4 object-based video coding. In particular, we propose two models to estimate the rate-distortion characteristics of coded objects as well as skipped objects. Based on the proposed models, we present several R-D coding modes with spatio-temporal trade-offs to improve coding efficiency. The proposed algorithm is performed at the object level for object-based video coding. Simulation results demonstrate moderate improvement at low as well as high bit rates. The proposed algorithm can produce the actual coded bits very close to the target bits over a wide range of bit rates. Consequently, the proposed algerian has not experienced any buffer overflow or underflow over the bit rates between 32 kbps and 256 kbps.

• Journal of KIISE:Information Networking
• /
• v.36 no.3
• /
• pp.198-203
• /
• 2009

With the growing demand for multimedia services and advances in display technology, new applications for 3$\sim$D scene communication have emerged. While multi-view video of these emerging applications may provide users with more realistic scene experience, drastic increase in the bandwidth is a major problem to solve. In this paper, we propose a fast prediction mode decision algorithm which can significantly reduce complexity and time consumption of the encoding process. This is based on the object segmentation, which can effectively identify the fast moving foreground object. As the foreground object with fast motion is more likely to be encoded in the view directional prediction mode, we can properly limit the motion compensated coding for a case in point. As a result, time savings of the proposed algorithm was up to average 45% without much loss in the quality of the image sequence.

• Journal of the Korean Institute of Telematics and Electronics S
• /
• v.34S no.5
• /
• pp.45-51
• /
• 1997

To trnsmit moving image signals by using an object vased coding technique at the rate of 8kbps~11kbps, it is very important to minimize the bit rates used for the compression of the color information of MF-objects. This paper proposes a hybrid coding method which uses the shpae adaptive coding method and the interframe reference method selectively. Gilge's shape adaptive orthogonal coidng method is utilized for shpae adaptive coding. The interfarame reference method approximates the low-passed signals of the image by gilge's shpae adaptive orthogonalization method and then refers the approximation error signals from the high frequency signal components of th eprevious frame. The proposed method achives the bit rates reduction of 17% compared to the gilge's shape adaptive orthogonalization method and 30% rduction compared to the shape adaptive DCT in average.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2021.06a
• /
• pp.43-46
• /
• 2021

최근 기계 임무수행에 사용되는 데이터양이 증가함에 따라 기계를 위한 효율적인 영상 압축방식의 필요성이 높아졌다. 기존의 비디오 코덱은 HVS (Human Visual System) 특성을 고려한 기술이기 때문에 부호화 과정에서 기계 임무수행에 필요하지 않은 정보를 효과적으로 제거할 수 없다. 반면 심층신경망 기반 압축네트워크의 경우, 원본 영상으로부터 기계 임무수행에 필수적인 데이터만을 추출하여 부호화 하도록 학습할 수 있는 장점이 있다. 본 논문에서는 압축 심층신경망과 기계 임무수행 네트워크로 구성되는 VCM (Video Coding for Machine) 프레임워크를 제안하고 학습에 의한 압축효율 향상을 검증한다. 이를 위해 압축 심층신경망을 객체탐지 임무수행 네트워크와 함께 학습시킨 결과, VVC (Versatile Video Coding) 대비 평균 61.16%의 BD-rate 감소가 확인되었다. 뿐만 아니라, 학습된 압축 심층신경망은 객체분할 임무수행에서도 VVC 대비 평균 58.43%의 BD-rate 감소를 보여 다중 기계 임무의 효율적 수행이 가능함을 확인할 수 있었다.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.40 no.1
• /
• pp.114-123
• /
• 2003

This paper proposes an efficient method for encoding the shape information of the object in the image. The polygonal approximation method is categorized into a loss coding method and is widely used for approximating object's shape information. The proposed method selects less number of vertices than IRM (iterated refinement method) or PVS (progressive vertex selection) when the maximum distortion is given, so reduces the bit-rates. The proposed method selects the vertices of a polygon with a simple and efficient method considering the rate-distortion sense. We construct the shape information coder, which shows the outstanding performance in the rate-distortion sense, based on the conventional progressive vertex selection method and the new vertex selection condition that we propose in this paper. Simulation results show that the proposed method has better performance than other conventional vertex selection methods in the tate-distortion sense.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2022.06a
• /
• pp.146-149
• /
• 2022

기존 RDO(Rate Distortion Optimization) 기반 압축 방식은 압축 성능에 초점을 두기 때문에 영상 내 인지 특성이 무시될 수 있다. 따라서 RoI(Region of Interest)을 기반으로 압축률을 조절하는 연구가 고안[1, 2, 3, 4] 되었으며, HVS(Human Visual System) 관점에서 영상 내 중요한 부분에 대해 더 높은 품질로 영상을 압축하는 연구가 대부분이다. 최근 인공지능 기술이 발전함에 따라 지능형 영상 분석에 대한 수요가 증가하고 있으며, 이에 따라 머신 비전을 위한 영상 부호화 및 효율적인 전송에 대한 필요성이 대두되고 있다. 본 논문에서는 VVC(Versatile Video Coding)의 dQP(delta Quantization Parameter)를 활용하여 RoI(Region of Interest) 기반압축 방법을 제안하고, 두가지의 RoI 추출 방식을 소개한다. Detectron2 Faster R-CNN X101-FPN [5]의 첫번째 탐지기를 통해 후보 영역 기반 RoI 을 추출하고, 두번째 탐지기를 통해 객체 기반 RoI 을 추출하여, 영상 내 객체 부분과 비객체 부분으로 나누어 서로 다른 압축률로 압축을 수행하였으며, 이에 따른 성능을 비교하고자 한다.