na li

na li


Research Focus & Metrics

Academic Career Length

6Years (2021-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
15 December 2021 in IEEE Transactions on Broadcasting
Verified
Journal Article
Closed access
NLna li
na li
Yun ZhangYun Zhang
Yun Zhang
CKC.-C. Jay Kuo
C.-C. Jay Kuo
In this work, we propose a high efficiency intra video coding based on data-driven transform, which is able to learn the source distributions of intra prediction residuals and improve the subsequent transform coding efficiency. Firstly, we model learning based transform design as an optimization problem of maximizing energy compaction or decorrelation. A data-driven Subspace Approximation with Adjusted Bias (Saab) transform is analyzed and compared with the mainstream Discrete Cosine Transform (DCT) on their energy compaction and decorrelation capabilities. Secondly, we propose a Saab transform based intra video coding framework with offline Saab transform learning. Meanwhile, intra mode dependent Saab transform is developed. Then, Rate-Distortion (RD) gain of Saab transform based intra video coding is theoretically and experimentally analyzed in detail. Finally, three strategies that apply the Saab transform to intra video coding are developed to improve the coding efficiency. Experimental results demonstrate that the proposed $8\times 8$ Saab transform based intra coding can achieve Bjønteggard Delta Bit Rate (BDBR) from −1.19% to −10.00% and −3.07% on average as compared with the mainstream $8\times 8$ DCT based intra coding. In case of variable size transform unit setting, the proposed algorithm achieves BDBR from −0.17% to −6.09% and −1.80% on average, which outperform the DCT based and the neural network based schemes.

28 June 2024 in IEEE Transactions on Circuits and Systems for Video Technology
Verified
Journal Article
Closed access
LZLinwei Zhu
Linwei Zhu
Yun ZhangYun Zhang
Yun Zhang
NLna li
na li
WWWenhui Wu
Wenhui Wu
SWShiqi Wang
Shiqi Wang
KWONG Tak Wu SamKWONG Tak Wu Sam
KWONG Tak Wu Sam
To further improve the performance of Versatile Video Coding (VVC), a neural network based multi-level in-loop filtering framework for luma and chroma is presented in this letter, which includes Reference pixel Level (RL), Coding tree unit Level (CL), and Frame Level (FL). The neural network based filters in these levels can be flexibly enabled. In RL, the coding performance upper bound is analyzed and asymmetric convolution is designed. In CL, the pixels located at the bottom and rightmost have been assigned greater weights for loss calculation during training. In addition, the co-located luma is adopted in CL and FL chroma filtering for guiding chroma enhancement due to the high correlation between them. For the architecture of neural network, two input channel fusion schemes are combined to enjoy both of their benefits. Extensive experimental results show that the proposed multi-level in-loop filtering method can achieve 6.87%, 32.8%, and 36.9% bit rate reductions on average for Y, U, and V components under all intra configuration, which outperforms the state-of-the-art works.

1 January 2026 in IEEE Transactions on Image Processing
Verified
Journal Article
Closed access
Yun ZhangYun Zhang
Yun Zhang
FCFeifan Chen
Feifan Chen
NLna li
na li
ZGZhiwei Guo
Zhiwei Guo
XWXu Wang
Xu Wang
KWONG Tak Wu SamKWONG Tak Wu Sam
KWONG Tak Wu Sam
Colored point cloud comprising geometry and attribute components is one of the mainstream representations enabling realistic and immersive 3D applications. To generate large-scale and denser colored point clouds, we propose a deep learning-based Joint Geometry and Attribute Up-sampling (JGAU) method, which learns to model both geometry and attribute patterns and leverages the spatial attribute correlation. Firstly, we establish and release a large-scale dataset for colored point cloud up-sampling, named SYSU-PCUD, which has 121 large-scale colored point clouds with diverse geometry and attribute complexities in six categories and four sampling rates. Secondly, to improve the quality of up-sampled point clouds, we propose a deep learning-based JGAU framework to up-sample the geometry and attribute jointly. It consists of a geometry upsampling network and an attribute up-sampling network, where the latter leverages the up-sampled auxiliary geometry to model neighborhood correlations of the attributes. Thirdly, we propose two coarse attribute up-sampling methods, Geometric Distance Weighted Attribute Interpolation (GDWAI) and Deep Learning-based Attribute Interpolation (DLAI), to generate coarsely upsampled attributes for each point. Then, we propose an attribute enhancement module to refine the up-sampled attributes and generate high quality point clouds by further exploiting intrinsic attribute and geometry patterns. Extensive experiments show that Peak Signal-to-Noise Ratio (PSNR) achieved by the proposed JGAU are 33.90 dB, 32.10 dB, 31.10 dB, and 30.39 dB when upsampling rates are 4×, 8×, 12×, and 16×, respectively. Compared to the state-of-the-art schemes, the JGAU achieves an average of 2.32 dB, 2.47 dB, 2.28 dB and 2.11 dB PSNR gains at four upsampling rates, respectively, which are significant.

5 August 2025 in IEEE Transactions on Circuits and Systems for Video Technology
Verified
Journal Article
Closed access
Yun ZhangYun Zhang
Yun Zhang
SZShisheng Zhang
Shisheng Zhang
NLna li
na li
CFChunling Fan
Chunling Fan
RHRaouf Hamzaoui
Raouf Hamzaoui
The Picture-Wise Just Noticeable Difference (PW-JND) represents the visibility threshold of human vision when viewing distorted images. The PW-JND plays an important role in perceptual image processing and compression. However, predicting the PW-JND is challenging due to its dependence on image content, viewing conditions, and the viewer. In this paper, we propose a visual perception-assisted deep PW-JND (VP-JND) prediction model for image compression that combines data-driven methods with the perceptual mechanisms of human vision. First, we identify a correlation between PW-JND and conventional pixel-wise JND. Based on this observation, we design the VP-JND model, consisting of a pixel-wise JND model, a deep binary classifier (VP-JNDnet) and a binary block search algorithm for refining predictions. VP-JNDnet exploits the pixel-wise JND map of the original image to predict whether a compressed image is perceptually lossless. In addition, the model incorporates visual importance of content and regions by using a mixed attention module and calculating perceptual loss during training. Experimental results show that VP-JND achieved an average precision of 94.82% and a mean absolute difference of 3.92 in predicting the JPEG quality factor corresponding to the PW-JND on the MCL-JCI dataset, outperforming state-of-the-art JND models. When applied to perceptual lossless image coding, the predicted PW-JND enabled average bit rate savings of 89.35% for JPEG compression on MCL-JCI and 85.46%/41.13% for JPEG/BPG compression on KonJND-1k. These savings were relative to images compressed at the lowest distortion level. The source codes and trained models are publicly available at https://github.com/SYSU-Video/VP-JND.

4 March 2022 in ACM Transactions on Multimedia Computing, Communications, and Applications
Verified
Journal Article
Open access
JLJinzhi Lin
Jinzhi Lin
Yun ZhangYun Zhang
Yun Zhang
NLna li
na li
HJHongling Jiang
Hongling Jiang
Ultra High-Definition (UHD) and Virtual Reality (VR) video streaming over 5G networks are emerging, in which High-Efficiency Video Coding (HEVC) is used as source coding to compress videos more efficiently and polar code is used as channel coding to transmit bitstream reliably over an error-prone channel. In this article, a novel Joint Source-Channel Decoding (JSCD) of polar codes for HEVC-based video streaming is presented to improve the streaming reliability and visual quality. Firstly, a Kernel Density Estimation (KDE) fitting approach is proposed to estimate the positions of error channel decoded bits. Secondly, a modified polar decoder called R-SCFlip is designed to improve the channel decoding accuracy. Finally, to combine the KDE estimator and the R-SCFlip decoder together, the JSCD scheme is implemented in an iterative process. Extensive experimental results reveal that, compared to the conventional methods without JSCD, the error data-frame correction ratios are increased. Averagely, 1.07% and 1.11% Frame Error Ratio (FER) improvements have been achieved for Additive White Gaussian Noise (AWGN) and Rayleigh fading channels, respectively. Meanwhile, the qualities of the recovered videos are significantly improved. For the 2D videos, the average Peak Signal-to-Noise Ratio (PSNR) and Structural SIMilarity (SSIM) gains reach 14% and 34%, respectively. For the 360֯ videos, the average improvements in terms of Weighted-to-Spherically-uniform PSNR (WS-PSNR) and Voronoi-based Video Multimethod Assessment Fusion (VI-VMAF) reach 21% and 7%, respectively.

18 December 2025 in IEEE Transactions on Multimedia
Verified
Journal Article
Closed access
SFShiyu Feng
Shiyu Feng
LZLinwei Zhu
Linwei Zhu
Yun ZhangYun Zhang
Yun Zhang
NLna li
na li
SWShiqi Wang
Shiqi Wang
Learned Image Compression (LIC) has achieved superior performance in recent years, of which the context entropy model is an important component. However, in the context entropy model, there is no deterministic correlation between neighboring channels, and it is difficult to capture inter-channel correlation as well as spatial correlation for further improving the performance. To address this issue, a Cubic-Checkerboard conTeXt entropy model (C-CTX) for LIC is proposed in this work, which is able to refer uniformly across the channel domain and maintain the correlations in the spatial domain. To make neighboring channels have more similar distribution, Cubic Checkerboard Mask (CCM) with channel- wise mask convolution is utilized to achieve uniform distribution in different domains and Channel Wise Re-Arrangement (CWRA) is performed in terms of entropy. Based on CCM and CWRA, two Feature Disentangle Modules (FDMs) are designed in C-CTX to project the context information within sub-spaces for catching spatial correlation and channel correlation separately. Extensive experimental evaluations show that our method outperforms the state-of-the-art works on six datasets, i.e., Kodak, Tecnick, CLIC'20, CLIC'21, CLIC'22, and JPEG-AI.