Xi Zhang
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
Xi Zhang has a total of 13 co-authors. Here are the top co-authors by number of shared publications:
- Xiaolin Wu
School of Computing and Artificial Intelligence, Southwest Jiaotong University, Chengdu, China
5 shared publications - Hao Xu
McMaster University
2 shared publications - Arpita Nema
Nanyang Technological University
1 shared publication - Hanwei Zhu
Nanyang Technological University
1 shared publication - Huihe Shao
School of Electronic Information and Electrical Engineering, Shanghai Jiao Tong University,Department of Automation,Shanghai,China,200240
1 shared publication - Kai Leung Yung
Department of Industrial and Systems Engineering, The Hong Kong Polytechnic University, Hong Kong, China
1 shared publication
Latest Publications
- 3 October 2026
- Published by Springer Nature
- p. 175-194
LongVQUBench: Benchmarking Long-Term Video Quality Understanding of Vision-Language Models
- 3 October 2026
- Published by Springer Nature
- p. 175-194
LongVQUBench: Benchmarking Long-Term Video Quality Understanding of Vision-Language Models
- 29 May 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Image Processing
- Vol. 35 (10577149) , 5864-5879
Improving 3D Gaussian Splatting Compression by Scene-Adaptive Lattice Vector Quantization
3D Gaussian Splatting (3DGS) is rapidly gaining popularity for its photorealistic rendering quality and real-time performance, but it generates massive amounts of data. Hence compressing 3DGS data is necessary for the cost effectiveness of 3DGS models. Recently, several anchor-based neural compression methods have been proposed, achieving good 3DGS compression performance. However, they all rely on uniform scalar quantization (USQ) due to its simplicity. A tantalizing question is whether more sophisticated quantizers can improve the current 3DGS compression methods with very little extra overhead and minimal change to the system. The answer is yes by replacing USQ with lattice vector quantization (LVQ). To better capture scene-specific characteristics, we optimize the lattice basis for each scene, improving LVQ’s adaptability and R-D efficiency. This scene-adaptive LVQ (SALVQ) strikes a balance between the R-D efficiency of vector quantization and the low complexity of USQ. SALVQ can be seamlessly integrated into existing 3DGS compression architectures, enhancing their R-D performance with minimal modifications and computational overhead. Moreover, by scaling the lattice basis vectors, SALVQ can dynamically adjust lattice density, enabling a single model to accommodate multiple bit rate targets. This flexibility eliminates the need to train separate models for different compression levels, significantly reducing training time and memory consumption.
- 29 May 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Image Processing
- Vol. 35 (10577149) , 5864-5879
Improving 3D Gaussian Splatting Compression by Scene-Adaptive Lattice Vector Quantization
3D Gaussian Splatting (3DGS) is rapidly gaining popularity for its photorealistic rendering quality and real-time performance, but it generates massive amounts of data. Hence compressing 3DGS data is necessary for the cost effectiveness of 3DGS models. Recently, several anchor-based neural compression methods have been proposed, achieving good 3DGS compression performance. However, they all rely on uniform scalar quantization (USQ) due to its simplicity. A tantalizing question is whether more sophisticated quantizers can improve the current 3DGS compression methods with very little extra overhead and minimal change to the system. The answer is yes by replacing USQ with lattice vector quantization (LVQ). To better capture scene-specific characteristics, we optimize the lattice basis for each scene, improving LVQ’s adaptability and R-D efficiency. This scene-adaptive LVQ (SALVQ) strikes a balance between the R-D efficiency of vector quantization and the low complexity of USQ. SALVQ can be seamlessly integrated into existing 3DGS compression architectures, enhancing their R-D performance with minimal modifications and computational overhead. Moreover, by scaling the lattice basis vectors, SALVQ can dynamically adjust lattice density, enabling a single model to accommodate multiple bit rate targets. This flexibility eliminates the need to train separate models for different compression levels, significantly reducing training time and memory consumption.
- 21 November 2024
- Published by Springer Nature
- p. 270-288
Fast Point Cloud Geometry Compression with Context-Based Residual Coding and INR-Based Refinement
- 21 November 2024
- Published by Springer Nature
- p. 270-288
Fast Point Cloud Geometry Compression with Context-Based Residual Coding and INR-Based Refinement
- 8 March 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Pattern Analysis and Machine Intelligence
- Vol. 45 (2) , 2024-2037
Multi-Modality Deep Restoration of Extremely Compressed Face Videos
Arguably the most common and salient object in daily video communications is the talking head, as encountered in social media, virtual classrooms, teleconferences, news broadcasting, talk shows, etc. When communication bandwidth is limited by network congestions or cost effectiveness, compression artifacts in talking head videos are inevitable. The resulting video quality degradation is highly visible and objectionable due to high acuity of human visual system to faces. To solve this problem, we develop a multi-modality deep convolutional neural network method for restoring face videos that are aggressively compressed. The main innovation is a new DCNN architecture that incorporates known priors of multiple modalities: the video-synchronized speech signal and semantic elements of the compression code stream, including motion vectors, code partition map and quantization parameters. These priors strongly correlate with the latent video and hence they are able to enhance the capability of deep learning to remove compression artifacts. Ample empirical evidences are presented to validate the superior performance of the proposed DCNN method on face videos over the existing state-of-the-art methods.
- 8 March 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Pattern Analysis and Machine Intelligence
- Vol. 45 (2) , 2024-2037
Multi-Modality Deep Restoration of Extremely Compressed Face Videos
Arguably the most common and salient object in daily video communications is the talking head, as encountered in social media, virtual classrooms, teleconferences, news broadcasting, talk shows, etc. When communication bandwidth is limited by network congestions or cost effectiveness, compression artifacts in talking head videos are inevitable. The resulting video quality degradation is highly visible and objectionable due to high acuity of human visual system to faces. To solve this problem, we develop a multi-modality deep convolutional neural network method for restoring face videos that are aggressively compressed. The main innovation is a new DCNN architecture that incorporates known priors of multiple modalities: the video-synchronized speech signal and semantic elements of the compression code stream, including motion vectors, code partition map and quantization parameters. These priors strongly correlate with the latent video and hence they are able to enhance the capability of deep learning to remove compression artifacts. Ample empirical evidences are presented to validate the superior performance of the proposed DCNN method on face videos over the existing state-of-the-art methods.
- Open Access
Attention-guided Image Compression by Deep Reconstruction of Compressive Sensed Saliency Skeleton
We propose a deep learning system for attention-guided dual-layer image compression (AGDL). In the AGDL compression system, an image is encoded into two layers, a base layer and an attention-guided refinement layer. Unlike the existing ROI image compression methods that spend an extra bit budget equally on all pixels in ROI, AGDL employs a CNN module to predict those pixels on and near a saliency sketch within ROI that are critical to perceptual quality. Only the critical pixels are further sampled by compressive sensing (CS) to form a very compact refinement layer. Another novel CNN method is developed to jointly decode the two compression layers for a much refined reconstruction, while strictly satisfying the transmitted CS constraints on perceptually critical pixels. Extensive experiments demonstrate that the proposed AGDL system advances the state of the art in perception-aware image compression.
Open AccessAttention-guided Image Compression by Deep Reconstruction of Compressive Sensed Saliency Skeleton
We propose a deep learning system for attention-guided dual-layer image compression (AGDL). In the AGDL compression system, an image is encoded into two layers, a base layer and an attention-guided refinement layer. Unlike the existing ROI image compression methods that spend an extra bit budget equally on all pixels in ROI, AGDL employs a CNN module to predict those pixels on and near a saliency sketch within ROI that are critical to perceptual quality. Only the critical pixels are further sampled by compressive sensing (CS) to form a very compact refinement layer. Another novel CNN method is developed to jointly decode the two compression layers for a much refined reconstruction, while strictly satisfying the transmitted CS constraints on perceptually critical pixels. Extensive experiments demonstrate that the proposed AGDL system advances the state of the art in perception-aware image compression.
- 2 December 2020
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Image Processing
- Vol. 30 (10577149) , 963-975
Ultra High Fidelity Deep Image Decompression With l∞-Constrained Compression
We propose a novel asymmetric image compression system of light ℓ ∞ -constrained predictive encoding and heavy-duty CNN-based soft decoding. The system achieves superior rate-distortion performances over the best of existing image compression methods, including BPG, WebP, FLIF and recent CNN codecs, in both ℓ 2 and ℓ ∞ error metrics, for bit rates near or above the threshold of perceptually transparent reconstruction. These remarkable coding gains are made by deep learning for compression artifact removal. A restoration CNN is designed to map a lossy compressed image to its original. Its unique strength is to enforce a tight error bound on a per pixel basis. As such, no small distinctive structures of the original image can be dropped or distorted, even if they are statistical outliers that are otherwise sacrificed by mainstream CNN restoration methods.
- 2 December 2020
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Image Processing
- Vol. 30 (10577149) , 963-975
Ultra High Fidelity Deep Image Decompression With l∞-Constrained Compression
We propose a novel asymmetric image compression system of light ℓ ∞ -constrained predictive encoding and heavy-duty CNN-based soft decoding. The system achieves superior rate-distortion performances over the best of existing image compression methods, including BPG, WebP, FLIF and recent CNN codecs, in both ℓ 2 and ℓ ∞ error metrics, for bit rates near or above the threshold of perceptually transparent reconstruction. These remarkable coding gains are made by deep learning for compression artifact removal. A restoration CNN is designed to map a lossy compressed image to its original. Its unique strength is to enforce a tight error bound on a per pixel basis. As such, no small distinctive structures of the original image can be dropped or distorted, even if they are statistical outliers that are otherwise sacrificed by mainstream CNN restoration methods.
- 6 February 2020
- Published by Springer Nature in The Journal of Supercomputing
- Vol. 76 (11) , 8922-8937
EDGAN: motion deblurring algorithm based on enhanced generative adversarial networks
- 6 February 2020
- Published by Springer Nature in The Journal of Supercomputing
- Vol. 76 (11) , 8922-8937
EDGAN: motion deblurring algorithm based on enhanced generative adversarial networks
- 1 December 2006
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. 3 (21511802) , 222-227
Nonlinear On-line Process Monitoring and Fault Detection Based on Kernel ICA
In this paper, a novel nonlinear process monitoring and fault detection method based on kernel ICA is proposed. The Kernel ICA method is a two-phase algorithm, KPCA first spheres data and makes the data structure become as linearly separable as possible using an implicit nonlinear mapping determined by kernel. Then ICA seeks the projection directions in the KPCA whitened space, making the distribution of the projected data as non-gaussian as possible. The application to the FCCU simulated process indicates that the proposed process monitoring method based on Kernel ICA can effectively capture the nonlinear relationship in process variables. Its performance significantly outperforms monitoring method based on ICA or KPCA.
- 1 December 2006
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. 3 (21511802) , 222-227
Nonlinear On-line Process Monitoring and Fault Detection Based on Kernel ICA
In this paper, a novel nonlinear process monitoring and fault detection method based on kernel ICA is proposed. The Kernel ICA method is a two-phase algorithm, KPCA first spheres data and makes the data structure become as linearly separable as possible using an implicit nonlinear mapping determined by kernel. Then ICA seeks the projection directions in the KPCA whitened space, making the distribution of the projected data as non-gaussian as possible. The application to the FCCU simulated process indicates that the proposed process monitoring method based on Kernel ICA can effectively capture the nonlinear relationship in process variables. Its performance significantly outperforms monitoring method based on ICA or KPCA.