Bozhong Liu
Bozhong Liu has a total of 17 co-authors. Here are the top co-authors by number of shared publications:
- Weidong Qiu
Shanghai Jiao Tong University
3 shared publications Ling Chen
University of Technology, Sydney
2 shared publicationsWeidong Qiu
Shanghai Jiao Tong University
2 shared publicationsCan Ge
School of Electronic Information and Electrical Engineering, Shanghai Jiao Tong University, Shanghai, China
1 shared publicationChengqi Zhang
University of Technology Sydney
1 shared publicationChunyang Liu
University of Technology Sydney
1 shared publication
Latest Publications
- Open Access
Adaptive Speech Quality Aware Complex Neural Network for Acoustic Echo Cancellation with Supervised Contrastive Learning
Acoustic echo cancellation (AEC) is designed to remove echoes, reverberation, and unwanted added sounds from the microphone signal while maintaining the quality of the near-end speaker's speech. This paper proposes adaptive speech quality complex neural networks to focus on specific tasks for real-time acoustic echo cancellation. In specific, we propose a complex modularize neural network with different stages to focus on feature extraction, acoustic separation, and mask optimization receptively. Furthermore, we adopt the contrastive learning framework and novel speech quality aware loss functions to further improve the performance. The model is trained with 72 hours for pre-training and then 72 hours for fine-tuning. The proposed model outperforms the state-of-the-art performance.
Open AccessAdaptive Speech Quality Aware Complex Neural Network for Acoustic Echo Cancellation with Supervised Contrastive Learning
Acoustic echo cancellation (AEC) is designed to remove echoes, reverberation, and unwanted added sounds from the microphone signal while maintaining the quality of the near-end speaker's speech. This paper proposes adaptive speech quality complex neural networks to focus on specific tasks for real-time acoustic echo cancellation. In specific, we propose a complex modularize neural network with different stages to focus on feature extraction, acoustic separation, and mask optimization receptively. Furthermore, we adopt the contrastive learning framework and novel speech quality aware loss functions to further improve the performance. The model is trained with 72 hours for pre-training and then 72 hours for fine-tuning. The proposed model outperforms the state-of-the-art performance.
- 7 September 2018
- Published by Springer Nature in Knowledge and Information Systems
- Vol. 60 (3) , 1307-1328
Encrypted data indexing for the secure outsourcing of spectral clustering
- 7 September 2018
- Published by Springer Nature in Knowledge and Information Systems
- Vol. 60 (3) , 1307-1328
Encrypted data indexing for the secure outsourcing of spectral clustering
- 1 March 2018
- Published by Springer Nature in Peer-to-Peer Networking and Applications
- Vol. 11 (2) , 252-264
Protecting lightweight block cipher implementation in mobile big data computing
- 1 March 2018
- Published by Springer Nature in Peer-to-Peer Networking and Applications
- Vol. 11 (2) , 252-264
Protecting lightweight block cipher implementation in mobile big data computing
- Open Access
- 1 January 2017
- Published by Institution of Engineering and Technology (IET) in Chinese Journal of Electronics
- Vol. 26 (1) , 93-100
Automatic Search of Threshold Implementations of 4‐Bit S‐Boxes Resisting DPA
The threshold implementation method of Substitution box (S-box) has been proposed by Nikova et al. for resisting first-order Differential power attacks with glitches. To lower the time complexity for a threshold implementation of a specific non-linear function, one needs to decompose the function first and then search possible share methods for it. However, the time complexity for this search process is still non-trivial. In this paper, an effective method of searching threshold implementations of 4-bit S-boxes is proposed. It mainly consists of two stages. For the decomposing stage, an efficient way of decomposing an S-box is introduced. For the sharing stage, the search complexity is lowered by the technique of time memory trade-off. As a result, threshold implementations of various lightweight block ciphers' S-boxes are given. Moreover, our method is applied to each 4-bit involutive S-box and some candidates of threshold implementations are presented.
- 1 January 2017
- Published by Institution of Engineering and Technology (IET) in Chinese Journal of Electronics
- Vol. 26 (1) , 93-100
Open AccessAutomatic Search of Threshold Implementations of 4‐Bit S‐Boxes Resisting DPA
The threshold implementation method of Substitution box (S-box) has been proposed by Nikova et al. for resisting first-order Differential power attacks with glitches. To lower the time complexity for a threshold implementation of a specific non-linear function, one needs to decompose the function first and then search possible share methods for it. However, the time complexity for this search process is still non-trivial. In this paper, an effective method of searching threshold implementations of 4-bit S-boxes is proposed. It mainly consists of two stages. For the decomposing stage, an efficient way of decomposing an S-box is introduced. For the sharing stage, the search complexity is lowered by the technique of time memory trade-off. As a result, threshold implementations of various lightweight block ciphers' S-boxes are given. Moreover, our method is applied to each 4-bit involutive S-box and some candidates of threshold implementations are presented.
- 17 September 2016
- Published by Springer Nature
- p. 66-78
Mining Co-locations from Continuously Distributed Uncertain Spatial Data
- 17 September 2016
- Published by Springer Nature
- p. 66-78
Mining Co-locations from Continuously Distributed Uncertain Spatial Data
- 1 November 2014
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. 2, 105-110
A New Approach to Multimedia Files Carving
Traditional file recovery methods rely on file system information, which are ineffective when file system information isn't available. File carving is a file recovery method that recovers files according to their structure and content without file system information, which is widely used in digital forensics. As the important carriers of digital information, multimedia files are important digital evidence. In this paper, a new multimedia file carving approach is proposed to improve the recovery accuracy of high entropy file fragments. The fragmented files can be recovered by a hierarchical carving process, including file header identification via entropy, file fragment type classification, and file reassembly via parallel unique path approach. A new file type classification method is constructed based on support vector machine, by using the features of BFD (byte frequency distribution) and ROC (rate of change). Four different datasets, such as DFRWS 2006/2007 challenge datasets, dataset simulating actual disk, dataset with randomly disordered fragments, and dataset with biomedical images, are employed in our experiments. The results show that JPEG recovery accuracy is improved greatly compared with that of Photo Rec tool. Our method performs best in the situation where the order of fragments is completely confusing.
- 1 November 2014
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. 2, 105-110
A New Approach to Multimedia Files Carving
Traditional file recovery methods rely on file system information, which are ineffective when file system information isn't available. File carving is a file recovery method that recovers files according to their structure and content without file system information, which is widely used in digital forensics. As the important carriers of digital information, multimedia files are important digital evidence. In this paper, a new multimedia file carving approach is proposed to improve the recovery accuracy of high entropy file fragments. The fragmented files can be recovered by a hierarchical carving process, including file header identification via entropy, file fragment type classification, and file reassembly via parallel unique path approach. A new file type classification method is constructed based on support vector machine, by using the features of BFD (byte frequency distribution) and ROC (rate of change). Four different datasets, such as DFRWS 2006/2007 challenge datasets, dataset simulating actual disk, dataset with randomly disordered fragments, and dataset with biomedical images, are employed in our experiments. The results show that JPEG recovery accuracy is improved greatly compared with that of Photo Rec tool. Our method performs best in the situation where the order of fragments is completely confusing.