Jongho Kim
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Jongho Kim has a total of 19 co-authors. Here are the top co-authors by number of shared publications:
- Jin Soo Choi
ETRI, Daejeon, Korea
5 shared publications - Byung-Gyu Kim
Embedded S/W Research Division, Electronics and Telecommunications Research Institute (ETRI), 161 Gajeong-dong, Yuseong-gu, Daejeon 305-350, Korea
3 shared publications - Chang-Sik Cho
Future Computing Research Division Artificial Intelligence Research Lab ETRI 34129 Daejeon South KOREA
3 shared publications - Se Yoon Jeong
Electronics and Telecommunications Research Institute, Daejeon, Korea (the Republic of)
3 shared publications - Byung-Gyu Kim
Sookmyung Women’s University
2 shared publications - Dongsan Jun
Department of Information and Communication Engineering, Kyungnam University, Changwon 51767, Korea
2 shared publications
Latest Publications
- Open Access
- 30 September 2024
- Published by Korea Multimedia Society - English Version Journal in Journal of Multimedia Information System
- Vol. 11 (3) , 175-184
Deep Inter Prediction for Versatile Video Coding (VVC)
한국멀티미디어학회(The Journal of Curriculum Evaluation)
- 30 September 2024
- Published by Korea Multimedia Society - English Version Journal in Journal of Multimedia Information System
- Vol. 11 (3) , 175-184
Open AccessDeep Inter Prediction for Versatile Video Coding (VVC)
한국멀티미디어학회(The Journal of Curriculum Evaluation)
- Open Access
Attention-Based Bi-Prediction Network for Versatile Video Coding (VVC) over 5G Network
As the demands of various network-dependent services such as Internet of things (IoT) applications, autonomous driving, and augmented and virtual reality (AR/VR) increase, the fifthgeneration (5G) network is expected to become a key communication technology. The latest video coding standard, versatile video coding (VVC), can contribute to providing high-quality services by achieving superior compression performance. In video coding, inter bi-prediction serves to improve the coding efficiency significantly by producing a precise fused prediction block. Although block-wise methods, such as bi-prediction with CU-level weight (BCW), are applied in VVC, it is still difficult for the linear fusion-based strategy to represent diverse pixel variations inside a block. In addition, a pixel-wise method called bi-directional optical flow (BDOF) has been proposed to refine bi-prediction block. However, the non-linear optical flow equation in BDOF mode is applied under assumptions, so this method is still unable to accurately compensate various kinds of bi-prediction blocks. In this paper, we propose an attention-based bi-prediction network (ABPN) to substitute for the whole existing bi-prediction methods. The proposed ABPN is designed to learn efficient representations of the fused features by utilizing an attention mechanism. Furthermore, the knowledge distillation (KD)- based approach is employed to compress the size of the proposed network while keeping comparable output as the large model. The proposed ABPN is integrated into the VTM-11.0 NNVC-1.0 standard reference software. When compared with VTM anchor, it is verified that the BD-rate reduction of the lightweighted ABPN can be up to 5.89% and 4.91% on Y component under random access (RA) and low delay B (LDB), respectively.
Open AccessAttention-Based Bi-Prediction Network for Versatile Video Coding (VVC) over 5G Network
As the demands of various network-dependent services such as Internet of things (IoT) applications, autonomous driving, and augmented and virtual reality (AR/VR) increase, the fifthgeneration (5G) network is expected to become a key communication technology. The latest video coding standard, versatile video coding (VVC), can contribute to providing high-quality services by achieving superior compression performance. In video coding, inter bi-prediction serves to improve the coding efficiency significantly by producing a precise fused prediction block. Although block-wise methods, such as bi-prediction with CU-level weight (BCW), are applied in VVC, it is still difficult for the linear fusion-based strategy to represent diverse pixel variations inside a block. In addition, a pixel-wise method called bi-directional optical flow (BDOF) has been proposed to refine bi-prediction block. However, the non-linear optical flow equation in BDOF mode is applied under assumptions, so this method is still unable to accurately compensate various kinds of bi-prediction blocks. In this paper, we propose an attention-based bi-prediction network (ABPN) to substitute for the whole existing bi-prediction methods. The proposed ABPN is designed to learn efficient representations of the fused features by utilizing an attention mechanism. Furthermore, the knowledge distillation (KD)- based approach is employed to compress the size of the proposed network while keeping comparable output as the large model. The proposed ABPN is integrated into the VTM-11.0 NNVC-1.0 standard reference software. When compared with VTM anchor, it is verified that the BD-rate reduction of the lightweighted ABPN can be up to 5.89% and 4.91% on Y component under random access (RA) and low delay B (LDB), respectively.
- Open Access
- 1 August 2014
- Published by Wiley in ETRI Journal
- Vol. 36 (4) , 528-536
Efficient Inter Prediction Mode Decision Method for Fast Motion Estimation in High Efficiency Video Coding
- 1 August 2014
- Published by Wiley in ETRI Journal
- Vol. 36 (4) , 528-536
Open AccessEfficient Inter Prediction Mode Decision Method for Fast Motion Estimation in High Efficiency Video Coding
- Open Access
- 1 October 2012
- Published by Wiley in ETRI Journal
- Vol. 34 (5) , 753-758
An SAD-Based Selective Bi-prediction Method for Fast Motion Estimation in High Efficiency Video Coding
- 1 October 2012
- Published by Wiley in ETRI Journal
- Vol. 34 (5) , 753-758
Open AccessAn SAD-Based Selective Bi-prediction Method for Fast Motion Estimation in High Efficiency Video Coding
- Open Access
- 7 April 2011
- Published by Wiley in ETRI Journal
- Vol. 33 (2) , 145-154
Highly Efficient Video Codec for Entertainment-Quality
- 7 April 2011
- Published by Wiley in ETRI Journal
- Vol. 33 (2) , 145-154
Open AccessHighly Efficient Video Codec for Entertainment-Quality
- 9 October 2007
- Published by Elsevier in Journal of Visual Communication and Image Representation
- Vol. 19 (3) , 175-183
Fast block mode decision algorithm in H.264/AVC video coding
- 9 October 2007
- Published by Elsevier in Journal of Visual Communication and Image Representation
- Vol. 19 (3) , 175-183
Fast block mode decision algorithm in H.264/AVC video coding
- 1 July 2007
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. 11 (19457871) , 1599-1602
Distortion-Based Partial Distortion Search for Fast Motion Estimation
Block motion estimation with full search is computationally complex. To reduce this complexity, different methods have been proposed, including partial distortion, which can reduce the computational complexity with no loss of image quality. We propose a distortion-based partial distortion search (DPDS) based on the magnitude of distortion and adaptive update of the matching order. We calculate absolute differences for all pixels in the predicted block point. Pixels are then sorted by the amount of distortion in a descending order for the matching process, which produces a scanning map. The sum of the absolute differences (SAD) of other candidate positions is then computed from this matching order. We also use an update of the scanning map by checking the increase in the number of absolute differences for the SAD value. The proposed DPDS algorithm improves the computational efficiency, compared with the original PDS scheme, because the accumulated value of the absolute pixel differences can rapidly reach the current minimum SAD value. The proposed algorithm is 4-13 times faster than the full search method with the same visual quality.
- 1 July 2007
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. 11 (19457871) , 1599-1602
Distortion-Based Partial Distortion Search for Fast Motion Estimation
Block motion estimation with full search is computationally complex. To reduce this complexity, different methods have been proposed, including partial distortion, which can reduce the computational complexity with no loss of image quality. We propose a distortion-based partial distortion search (DPDS) based on the magnitude of distortion and adaptive update of the matching order. We calculate absolute differences for all pixels in the predicted block point. Pixels are then sorted by the amount of distortion in a descending order for the matching process, which produces a scanning map. The sum of the absolute differences (SAD) of other candidate positions is then computed from this matching order. We also use an update of the scanning map by checking the increase in the number of absolute differences for the SAD value. The proposed DPDS algorithm improves the computational efficiency, compared with the original PDS scheme, because the accumulated value of the absolute pixel differences can rapidly reach the current minimum SAD value. The proposed algorithm is 4-13 times faster than the full search method with the same visual quality.
- 1 June 2007
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- No. 0747668X,p. 1-5
A Fast Block Mode Decision Algorithm using a Block Correlation in H.264/AVC
The recent video coding standard H.264/AVC show extremely higher coding efficiency compare to any other previous standards. Although H.264/AVC can achieve over 50% of bit rate saving with same quality using the rate-distortion process, but it brings high computational complexity. In this paper, we propose an algorithm that can reduce the complexity of the codec by reducing the block mode decision process adaptively. In this paper an efficient method is proposed to reduce the inter mode decision coding complexity using the direct prediction methods based on inter block correlation and adaptive rate distortion cost threshold for early stopping. The intra mode decision complexity is also reduced using the early intra skip method. The experimental results show that the proposed algorithm can achieve up to 43.85~74.12% speed up ratio with a little PSNR loss. Increment in bit requirement is also not much noticeable.
- 1 June 2007
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- No. 0747668X,p. 1-5
A Fast Block Mode Decision Algorithm using a Block Correlation in H.264/AVC
The recent video coding standard H.264/AVC show extremely higher coding efficiency compare to any other previous standards. Although H.264/AVC can achieve over 50% of bit rate saving with same quality using the rate-distortion process, but it brings high computational complexity. In this paper, we propose an algorithm that can reduce the complexity of the codec by reducing the block mode decision process adaptively. In this paper an efficient method is proposed to reduce the inter mode decision coding complexity using the direct prediction methods based on inter block correlation and adaptive rate distortion cost threshold for early stopping. The intra mode decision complexity is also reduced using the early intra skip method. The experimental results show that the proposed algorithm can achieve up to 43.85~74.12% speed up ratio with a little PSNR loss. Increment in bit requirement is also not much noticeable.
- 1 October 2006
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- No. 15224880,p. 2357-2360
A Directional & Adaptive Diamond Search by Adaptive Pattern Switching with a Predicted Motion Vector
We propose a simple fast block-matching algorithm (BMA) based on the direction of the predicted motion vector called directional & adaptive diamond search by adaptive pattern switching (DADS-APS). The proposal method has two sequential search steps, including 1) an initial search, and 2) a refinement search for the local area. Adaptive pattern switching (APS) is proposed for the initial search and a unit-size rood pattern is used for the refinement. The initial search step consists of pattern size determination and selection of a pattern shape. We use an adaptive pattern size that is adjusted by the amount of motion. This method is superior to a fixed-pattern size algorithm, regardless of the amount of motion. In video sequences, each motion has a unique direction. Using this property, we use adaptive pattern switching between ARPS and DADS. APS considers the motion direction so we can easily and correctly find the minimal matching error (MME) point with less error distortion. APS can thus reduce the number of poorly related search points. Analysis shows that DADS-APS exhibits good PSNR performance and the average number of search points compares favorably with other methods.
- 1 October 2006
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- No. 15224880,p. 2357-2360
A Directional & Adaptive Diamond Search by Adaptive Pattern Switching with a Predicted Motion Vector
We propose a simple fast block-matching algorithm (BMA) based on the direction of the predicted motion vector called directional & adaptive diamond search by adaptive pattern switching (DADS-APS). The proposal method has two sequential search steps, including 1) an initial search, and 2) a refinement search for the local area. Adaptive pattern switching (APS) is proposed for the initial search and a unit-size rood pattern is used for the refinement. The initial search step consists of pattern size determination and selection of a pattern shape. We use an adaptive pattern size that is adjusted by the amount of motion. This method is superior to a fixed-pattern size algorithm, regardless of the amount of motion. In video sequences, each motion has a unique direction. Using this property, we use adaptive pattern switching between ARPS and DADS. APS considers the motion direction so we can easily and correctly find the minimal matching error (MME) point with less error distortion. APS can thus reduce the number of poorly related search points. Analysis shows that DADS-APS exhibits good PSNR performance and the average number of search points compares favorably with other methods.