A5040889357- Advanced Steganography and Watermarking Techniques
- Advanced Neural Network Applications
- Video Surveillance and Tracking Methods
- Advanced Image and Video Retrieval Techniques
- Image Processing and 3D Reconstruction
- Chaos-based Image/Signal Encryption
- Digital Media Forensic Detection
- 3D Shape Modeling and Analysis
- Computer Graphics and Visualization Techniques
- Face recognition and analysis
- Security in Wireless Sensor Networks
- Domain Adaptation and Few-Shot Learning
- Generative Adversarial Networks and Image Synthesis
- Image Enhancement Techniques
- Bone health and treatments
- Algorithms and Data Compression
- Color perception and design
- Remote Sensing and LiDAR Applications
- Grouting, Rheology, and Soil Mechanics
- Gait Recognition and Analysis
- Molecular Biology Techniques and Applications
- Advanced Image Fusion Techniques
- Complexity and Algorithms in Graphs
- Caching and Content Delivery
- Spectroscopy Techniques in Biomedical and Chemical Research
Yueyang Hospital
2025
Shanghai University of Traditional Chinese Medicine
2025
Guangzhou University
2015-2024
Mianyang Central Hospital
2023-2024
Chengdu University of Technology
2023
Guangxi University of Finance and Economics
2018-2020
Beth Israel Deaconess Medical Center
2016
Harvard University
2016
Beijing Forestry University
2016
Guangxi University
2011
Abstract Background Data based clinical decision support system is a boon for health care monitoring. Smart healthcare monitoring systems play vital role in the early diagnosis and detection of physical mental patients. The smart IoT (C-IoT) are data-driven provide efficient this purpose. Purpose There need to have secure, accurate, HCM that capable processing large amounts patient data timely various complications. Traditional ways migration imprecise, less do not cover all angles necessary...
Nickel (Ni) is a known carcinogen, although the mechanism of its carcinogenicity not clear. Here, we provide evidence that Ni can induce phosphorylation histone H3 at serine 10 residue in c-jun N-terminal kinase (JNK)/stress-activated protein (SAPK)-dependent manner. induces JNK, with no effect on states extracellular signal-regulated (ERK) or p38 mitogen-activated kinases. An inhibitor JNK eliminated Ni-initiated JNK-mediated induction 10, whereas inhibitors specific for ERK kinases had...
Breast cancer is the most commonly diagnosed type worldwide. Overexpression of human epidermal growth factor receptor 2 (HER2) an important subtype breast and results in increased risk recurrence metastasis patients. At present, immunohistochemistry (IHC) used to detect expression HER2 tissues as golden standard. However, IHC has some shortcomings, such large subjective impact, long time consumption, expensive reagents, etc. In this paper, a combined morphological spectroscopic diagnostic...
Mineral Prospectivity Mapping (MPM) plays a pivotal role in identifying geo-anomalies that are indicative of potential mineralization, drawing upon various geological, geophysical, geochemical, and remote sensing data. With the increasing availability such data recent years, data-driven MPM methods have proven effective discovering new mineral deposits, especially when utilizing machine learning techniques to unveil complex relationships between exploration occurrences. Deep learning,...
Visible-infrared cross-modality person re-identification (VI-ReID), whose aim is to match images between visible and infrared modality, a challenging image retrieval task. Most existing works integrate batch normalization layers into their neural network, but we found out that would lead two types of distribution gap: 1) inter-mini-batch gap- the gap same modality each mini-batch; 2) intra-mini-batch gap-the different within mini-batch. To address these problems, propose new layer called...
Visible-Infrared cross-modality person reidentification (VI-ReID), whose aim is to match images between visible and infrared modality, a challenging image retrieval task. Batch Hard Triplet loss widely used in re-identification tasks, but it does not perform well the Because only optimizes hardest triplet for each anchor within mini-batch, samples may all belong same which will lead imbalance problem of modality optimization. To address this problem, we adopt batch selection strategy,...
In this paper, we propose a novel watermarking-based secure image acquisition scheme for multimedia sensor networks (MSNs), which provides support authentication and restoration. the proposed scheme, node groups frames. Two successive frames compose non-overlapping restoration group. The watermarking bits, are version of down sampling image, computed from first reversibly embedded into second image. sink does verification by reversible watermark. Compared to previous work, our approach can...
Deep learning that jointly studies and extracts features is very promising for steganalysis. In this article, we design a simple but effective Multi-column Convolutional Neural Network (MCNN) based on steganalysis architecture images. The proposed MCNN allows the input image to be of arbitrary size or resolution. particular, by utilizing filters with receptive fields different sizes, learned each column CNN are adaptive variations in payloads. Comprehensive experiments standard dataset show...
As one of the most common cancers, accurate, rapid, and simple histopathological diagnosis is very important for breast cancer. Raman imaging a powerful technique label-free analysis tissue composition histopathology, but it suffers from slow speed when applied to large-area sections. In this study, we propose dual-modal method that combines mapping data with microscopy bright-field images achieve virtual staining cancer We validate our on various sections different morphologies biomarker...
According to the characteristics of multi-core architectures and binary storage property integer sequence, this paper proposes an efficient thread-level parallel algorithm for sorting sequence on computers. The divides input several data blocks in main memory distributes these into shared L2 cache private L1 respectively, implements dynamically load balance among processing cores, utilizes data-level SIMD instructions thread-binding technique speed up procedure. Experiment results show that...
This paper presents a robust spatial watermarking scheme for 3D point cloud models using self-similarity partition. In the new scheme, model is uniquely partitioned into patches octree structure and PCA preprocessing. Then concise similarity measurement designed to identify cluster similar patch chains codebook formed. By altering local vector direction of certain points each patch, watermark bits are embedded average every patch. The can be extracted based on three keys, known center,...
Abstract In Modification Direction Exploiting algorithm (EMD) exists binary information and multi-band data conversion redundancy problem, this paper proposed a new improved parallel algorithm. paper, we change the original EMD imbed function, introduce random safety parameter to improve security of When 2n is equal an integer power 2, does not have it can achieve maximum theory embedding rate efficiency. This utilizes multi-core computing operational Therefore, application more flexible efficient.
In this paper, the ANSYS finite element analysis is applied to pine backrest chair models. During structure design progress, Taguchi method for robust applied. By means of signal-noise ratio (SNR) factor effect analysis, optimal scheme structural quality determined that designed as trapezoidal seat frame, upper waling, bottom waling with II type and leaning back rear leg. Based on above condition, maximum equivalent stress improved by 9.05% than initial one, deformation ameliorated 14.42%...
This paper presents a new adaptive, high-capacity steganography for 3D point cloud models using self-similarity segmentation. Every embedding vertex of the model can adaptively embed variable σ (σ ≥ 3) bits an adaptive space subdivision procedure which uses normal direction vertexes to estimate capacity every with respect human visual system. The scheme segments patches measures. representative and similar are then taken as reference message patches. Finally, in has point-to-point...