Li Li

ORCID: 0000-0003-4799-5896
Publications
Citations
Views
---
Saved
---
About
Contact & Profiles
Research Areas
  • Adversarial Robustness in Machine Learning
  • Advanced Steganography and Watermarking Techniques
  • Digital Media Forensic Detection
  • Chaos-based Image/Signal Encryption
  • Anomaly Detection Techniques and Applications
  • Domain Adaptation and Few-Shot Learning
  • Advanced Malware Detection Techniques
  • Advanced Neural Network Applications
  • Generative Adversarial Networks and Image Synthesis
  • Integrated Circuits and Semiconductor Failure Analysis
  • Advanced Computational Techniques and Applications
  • Image and Video Stabilization
  • Software Testing and Debugging Techniques
  • Logic, programming, and type systems
  • Machine Learning in Materials Science
  • Stroke Rehabilitation and Recovery
  • Internet Traffic Analysis and Secure E-voting
  • Research studies in Vietnam
  • Military Defense Systems Analysis
  • Biometric Identification and Security
  • Machine Learning and Data Classification
  • Advanced Decision-Making Techniques
  • Bacillus and Francisella bacterial research
  • Distributed and Parallel Computing Systems
  • Machine Learning and Algorithms

Shanghai University
2010-2025

Tongji University
2020-2024

Beihang University
2023

University of Science and Technology of China
2001-2022

George Washington University
2021

Google (United States)
2021

Hefei University of Technology
2019

University Town of Shenzhen
2019

Tsinghua University
2019

North China Electric Power University
2008

Steganography is an information hiding technique for covert communication. So far Syndrome-Trellis Codes (STC), a convolutional codes-based method, the only near-optimal coding i.e., it can approach rate-distortion bound of content-adaptive steganography in practice. However, as secure communication application, needs diversity methods. This paper proposes another and better steganographic method based on polar codes, using Successive Cancellation List (SCL) decoding algorithm to minimize...

10.1109/tcomm.2020.2982624 article EN IEEE Transactions on Communications 2020-03-24

Traditional image steganography conceals secret messages in unprocessed natural images by modifying the pixel value, causing obtained stego to be different from original terms of statistical distribution; thereby, it can detected a well-trained classifier for steganalysis. To ensure is imperceptible and line with trend art produced Artificial-Intelligence-Generated Content (AIGC) becoming popular on social networks, this paper proposes embed hidden information throughout process generation...

10.3390/math12040615 article EN cc-by Mathematics 2024-02-19

Video watermarking based on frequency domain is proved to have good invisibility and robustness. However, most of the existing video schemes embed watermarks in subblock segmentation, while ignoring variation relationship between space features. Therefore, it difficult achieve robust authentication complex application scenarios. In this paper, a ring subband constructed DT-CWT as watermark embedding region by analyzing characteristics under multiple attacks. Subsequently, double embedded...

10.1109/tcsvt.2024.3405558 article EN IEEE Transactions on Circuits and Systems for Video Technology 2024-05-27

10.1109/jiot.2024.3434656 article EN IEEE Internet of Things Journal 2024-07-29

The rapid advancement of AI-generated content (AIGC) has significantly improved the realism and accessibility synthetic images. While large image generation models offer immense potential in creative industries, they also introduce serious challenges, including copyright infringement, authentication, traceability generated Digital watermarking emerged as a promising approach to address these concerns by embedding imperceptible yet detectable signatures into This survey provides comprehensive...

10.3390/math13040651 article EN cc-by Mathematics 2025-02-16

Mobile Multimedia Broadcasting (MMB) is a new hotspot of wireless application and Cloud computing shared services architecture based method. With the analysis on advantages disadvantages characteristics in traditional service modes MMB, method virtual distributed processing introduced to promote integration MMB industrial chain. The reasonable developing orientation involved several technologies as Computing, SOA, ubiquitous sensing network neural put forward from point view MMB.

10.1109/icmult.2010.5630979 article EN 2010-10-01

Existing methods for Deep Neural Networks (DNN) watermarking either require accessing the internal parameters of DNN models (white-box watermarking), or rely on backdooring to enforce a desired behavior model when is fed with specific set key input images (black-box watermarking). In this letter, we propose black-box multi-bit algorithm, suitable multiclass classification networks, whereby presence watermark can be retrieved from output network in correspondence <italic...

10.1109/lsp.2023.3239737 article EN IEEE Signal Processing Letters 2023-01-01

More and more edge devices mobile apps are leveraging deep learning (DL) capabilities. Deploying such models on – referred to as on-device rather than remote cloud-hosted services, has gained popularity because it avoids transmitting user's data off of the device achieves high response time. However, can be easily attacked, they accessed by unpacking corresponding model is fully exposed attackers. Recent studies show that attackers generate white-box-like attacks for an or even inverse its...

10.1145/3597926.3598113 preprint EN 2023-07-12

Side channel steganalysis refers to detecting a steganographer in social websites via behavior analysis. In this paper, we first design side based on the correlation between image sequences of users, which aims find out behaviorally anomalous steganographer. According experimental results steganalysis, it is intuitively secure for act identically normal users since she can avoid being detected by steganalysis. However, when faced with various detection methods, still behave similar user? To...

10.1109/tmm.2019.2959909 article EN IEEE Transactions on Multimedia 2019-12-17

Neural networks especially the convolution neural (CNN) have become prevalent and numerous CNN accelerators been developed to achieve higher performance. While clock frequency determines operation speed has direct influence on performance of accelerators, we propose apply overclocking, a circuit optimization approach that enables frequency, general accelerators. This technique brings significant improvement, but it leads moderate timing errors, wrong computing results low prediction...

10.1109/itc-asia.2019.00039 article EN 2019-09-01

This paper presents a DNN bottleneck reinforcement scheme to alleviate the vulnerability of Deep Neural Networks (DNN) against adversarial attacks. Typical classifiers encode input image into compressed latent representation more suitable for inference. information makes trade-off between image-specific structure and class-specific in an image. By reinforcing former while maintaining latter, any redundant information, be it or not, should removed from representation. Hence, this proposes...

10.1145/3394171.3413604 preprint EN Proceedings of the 30th ACM International Conference on Multimedia 2020-10-12

Deep learning (DL) has recently achieved tremendous success in a variety of cutting-edge applications, e.g., image recognition, speech and natural language processing, autonomous driving. Besides the available big data hardware evolution, DL frameworks platforms play key role to catalyze research, development, deployment intelligent solutions. However, difference computation paradigm, architecture design implementation existing brings challenges for software deployment, maintenance,...

10.48550/arxiv.1811.05187 preprint EN other-oa arXiv (Cornell University) 2018-01-01

Recently, adversarial attack methods have been developed to challenge the robustness of machine learning models. However, mainstream evaluation criteria experience limitations, even yielding discrepancies among results under different settings. By examining various algorithms, including gradient-based and query-based attacks, we notice lack a consensus on uniform standard for unbiased performance evaluation. Accordingly, propose Piece-wise Sampling Curving (PSC) toolkit effectively address...

10.48550/arxiv.2104.11103 preprint EN other-oa arXiv (Cornell University) 2021-01-01

10.1007/bf02948854 article EN Journal of Computer Science and Technology 2001-01-01

<title>Abstract</title> With the rapid development of artificial intelligence technology, more and deep-faked audios have emerged in cyberspace, leading to an urgent need for detecting tracing deep fake audios. Watermarking provides active method audios, especially through multi-bit watermarking technology. Multiple watermarks can provide stronger protection tracking capabilities various scenarios. Especially users upload that incorporate with both personal lyricist composer platform...

10.21203/rs.3.rs-5429669/v1 preprint EN cc-by Research Square (Research Square) 2024-11-29
Coming Soon ...