Zhaohui Cai

Zhaohui Cai

Auto-generated profile, based on publicly available information

Zhaohui Cai

School of Computer Science, Wuhan University, Wuhan, China

Research Focus & Metrics

Academic Career Length

19Years (2008-Present)

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

Representative Publications

View All Publications
2 August 2023 in IEEE Transactions on Image Processing
Verified
Journal Article
Closed access
QYQu Yang
Qu Yang
Mang YeMang Ye
Mang Ye
ZCZhaohui Cai
Zhaohui Cai
KSKehua Su
Kehua Su
BDBo Du
Bo Du
Composing Text and Image to Image Retrieval (CTI-IR) aims at finding the target image, which matches the query image visually along with the query text semantically. However, existing works ignore the fact that the reference text usually serves multiple functions, e.g., modification and auxiliary. To address this issue, we put forth a unified solution, namely Hierarchical Aggregation Transformer incorporated with Cross Relation Network (CRN). CRN unifies modification and relevance manner in a single framework. This configuration shows broader applicability, enabling us to model both modification and auxiliary text or their combination in triplet relationships simultaneously. Specifically, CRN includes: 1) Cross Relation Network comprehensively captures the relationships of various composed retrieval scenarios caused by two different query text types, allowing a unified retrieval model to designate adaptive combination strategies for flexible applicability; 2) Hierarchical Aggregation Transformer aggregates top-down features with Multi-layer Perceptron (MLP) to overcome the limitations of edge information loss in a window-based multi-stage Transformer. Extensive experiments demonstrate the superiority of the proposed CRN over all three fashion-domain datasets. Code is available at github.com/yan9qu/crn.

31 January 2018 in Sensors
Verified
Journal Article
Open access
PDF file
JWJun Wang
Jun Wang
RZRongbo Zhu
Rongbo Zhu
SLShubo Liu
Shubo Liu
ZCZhaohui Cai
Zhaohui Cai
Wireless sensor networks (WSNs) are widely applied in industrial application with the rapid development of Industry 4.0. Combining with centralized cloud platform, the enormous computational power is provided for data analysis, such as strategy control and policy making. However, the data analysis and mining will bring the issue of privacy leakage since sensors will collect varieties of data including sensitive location information of monitored objects. Differential privacy is a novel technique that can prevent compromising single record benefits. Geospatial data can be indexed by a tree structure; however, existing differentially private release methods pay no attention to the concrete analysis about the partition granularity of data domains. Based on the overall analysis of noise error and non-uniformity error, this paper proposes a data domain partitioning model, which is more accurate to choose the grid size. A uniform grid release method is put forward based on this model. In order to further reduce the errors, similar cells are merged, and then noise is added into the merged cells. Results show that our method significantly improves the query accuracy compared with other existing methods.

28 June 2023 in Algorithms
Verified
Journal Article
Open access
PDF file
JZJiachen Zhang
Jiachen Zhang
GTGuoqing Tu
Guoqing Tu
SLShubo Liu
Shubo Liu
ZCZhaohui Cai
Zhaohui Cai
The rapid development of speech synthesis technology has significantly improved the naturalness and human-likeness of synthetic speech. As the technical barriers for speech synthesis are rapidly lowering, the number of illegal activities such as fraud and extortion is increasing, posing a significant threat to authentication systems, such as automatic speaker verification. This paper proposes an end-to-end speech synthesis detection model based on audio feature fusion in response to the constantly evolving synthesis techniques and to improve the accuracy of detecting synthetic speech. The model uses a pre-trained wav2vec2 model to extract features from raw waveforms and utilizes an audio feature fusion module for back-end classification. The audio feature fusion module aims to improve the model accuracy by adequately utilizing the audio features extracted from the front end and fusing the information from timeframes and feature dimensions. Data augmentation techniques are also used to enhance the performance generalization of the model. The model is trained on the training and development sets of the logical access (LA) dataset of the ASVspoof 2019 Challenge, an international standard, and is tested on the logical access (LA) and deep-fake (DF) evaluation datasets of the ASVspoof 2021 Challenge. The equal error rate (EER) on ASVspoof 2021 LA and ASVspoof 2021 DF are 1.18% and 2.62%, respectively, achieving the best results on the DF dataset.

18 May 2023 in The Journal of Supercomputing
Verified
Journal Article
Closed access
PDF file
ZCZhuoqing Chang
Zhuoqing Chang
SLShubo Liu
Shubo Liu
RQRun Qiu
Run Qiu
SSSong Song
Song Song
ZCZhaohui Cai
Zhaohui Cai
GTGuoqing Tu
Guoqing Tu

6 January 2023 in The Journal of Supercomputing
Verified
Journal Article
Closed access
PDF file
JLJie Luo
Jie Luo
SLShubo Liu
Shubo Liu
ZCZhaohui Cai
Zhaohui Cai
CXChang Xiong
Chang Xiong
GTGuoqing Tu
Guoqing Tu

18 July 2023 in IEEE Transactions on Instrumentation and Measurement
Verified
Journal Article
Closed access
CXChang Xiong
Chang Xiong
ZCZhaohui Cai
Zhaohui Cai
SLShubo Liu
Shubo Liu
JLJie Luo
Jie Luo
GTGuoqing Tu
Guoqing Tu
Nonintrusive load monitoring (NILM) is a process that monitors the aggregated power consumption data of customers measured by a single sensor and decomposes the real-time power consumption of each dedicated device. Recent research has been focused on in-depth learning. However, the expensive cost of training time and the huge model scale are not conducive to the realization of smart meters. It is still challenging to accurately predict the real-time power of high-frequency appliances. This paper uses the discrete wavelet transform (DWT) to preprocess the data, which divides the frequency of the training power data. The processed data will be transferred to the enhanced sequence-to-point network (en-S2P) for training. The en-S2P network is the promotion of a sequence-to-point model (S2P) optimized by the autoencoder and bidirectional long short-term memory (Bi-LSTM) layer. And it is compressed by a combined pruning algorithm after being trained. The proposed model method is tested on the UK-DALE and REDD datasets using db2 and sym5 wavelet transform and compared with the S2P. Extensive experiments show that the proposed model can obtain a more satisfactory decomposition performance of appliances, especially high-frequency. In addition, it has a more lightweight scale and maintains a certain degree of sparsity, which is friendly to electricity meters.