Watermarking Deep Neural Networks in Image Processing.

Publishing/sharing pretrained deep neural network (DNN) models is a common practice in the community of computer vision. The increasing popularity of pretrained models has made it a serious concern: how to protect the intellectual properties of model owners and avert illegal usages by malicious attackers. This article aims at developing a framework for watermarking DNNs, with a particular focus on low-level image processing tasks that map images to images. Using image denoising and superresolution as case studies, we develop a black-box watermarking method for pretrained models, which exploits the overparameterization of the DNNs in image processing. In addition, an auxiliary module for visualizing the watermark information is proposed for further verification. Extensive experiments show that the proposed watermarking framework has no noticeable impact on model performance and enjoys the robustness against the often-seen attacks.

Tablets of paliperidone using compression-coated technology for controlled ascending release.

The aim of this work was to prepare ascending release compression-coated (CC) tablets with paliperidone (PAL) using a simple manufacturing technique and short manufacturing process. The release behavior and mechanisms in vitro of the final tablets was investigated and evaluated. The PAL CC tablets were comprised of a core layer of high viscosity hydroxypropyl cellulose (HPC-H) and a coating layer of high viscosity hydroxypropyl methylcellulose (HPMC-K100M). Several factors such as materials and core tablet compositions were studied for their influence in the formulation procedure. The drug release mechanism was studied using gravimetric analysis. The data could be fitted to the Peppas model. The ascending drug release results were expressed in terms of the slope of the release curve at different time points. Results showed that the formulation could achieve a good ascending drug release when the weight ratio of PAL was 5:1 (core:layer). The fraction of HPC and HPMC was 33 %, and the combination of Eudragit RL-PO was 10%. The ascending release mechanism was due to solvent penetration into the PAL CC tablets, and subsequent drug dissolution from the gelatinous HPC and HPMC matrix erosion. The release mechanism was therefore a combination of diffusion and erosion. This work demonstrated that the compression-coated tablets could achieve controlled ascending release over 24 h for the oral administration systems.

Chimeric peptide constructs comprising linear B-cell epitopes: application to the serodiagnosis of infectious diseases.

Linear B-cell epitopes are ideal biomarkers for the serodiagnosis of infectious diseases. However, the long-predicted diagnostic value of epitopes has not been realized. Here, we demonstrated a method, diagnostic epitopes in four steps (DEIFS), that delivers a combination of epitopes for the serodiagnosis of infectious diseases with a high success rate. Using DEIFS for malaria, we identified 6 epitopes from 8 peptides and combined them into 3 chimeric peptide constructs. Along with 4 other peptides, we developed a rapid diagnostic test (RDT), which is able to differentiate Plasmodium falciparum (P. falciparum) from Plasmodium vivax (P. vivax) infections with 95.6% overall sensitivity and 99.1% overall specificity. In addition to applications in diagnosis, DEIFS could also be used in the diagnosis of virus and bacterium infections, discovery of vaccine candidates, evaluation of vaccine potency, and study of disease progression.

Autoantibodies to chromogranin A are potential diagnostic biomarkers for non-small cell lung cancer.

Cancer-associated autoantibodies show promise as sensitive biomarkers for the early detection of cancer. To test the immunogenicity of chromogranin A (ChgA) as a B cell autoantigen and to assess the potential applications of ChgA autoantibodies as novel biomarkers for the diagnosis of non-small cell lung cancer (NSCLC), we developed a high-content peptide microarray using ChgA peptides. Autoantibody profiling was carried out using sera from 168 individuals with NSCLC and 97 healthy controls. We present evidence for the occurrence of autoantibodies to ChgA peptides in patient sera and identified five highly responsive peptides in the NSCLC group using significance analysis of microarray (SAM). Receiver operating characteristic analyses showed that ChgA autoantibodies are valuable in the predictive diagnosis of NSCLC, suggesting that serum autoantibodies to ChgA-derived peptides are promising novel markers of NSCLC. Moreover, the high-content peptide microarray antibody profiling reported in this work provides a powerful tool to visualize the overall B cell response to ChgA peptides and should enable the rapid development of in-depth research into ChgA.

Sustained-release pellets of nifedipine using microcrystals combined with MCC-based matrix.

The purpose of this study was to prepare sustained-release pellets of nifedipine (NSPs) based on MCC matrix. Wet-milling and extrusion-spheronization techniques were employed to prepare the microcrystals and pellets, respectively. The drug release mechanism and the influencing factors were investigated. After milled with HPMC (E5), the mean particle size of nifedipine in co-grinding mixture (CGM) was 5 µm, which is 15-fold smaller than that of raw material. DSC, X-ray powder diffraction and microscopic observation confirmed the microcrystals of drug were maintained in the CGM. With increased milling time and the content of HPMC, the dissolution rate was greatly enhanced compared with the raw material. The NSPs prepared by MCC and the CGM, which was obtained by cogrinding nifedipine with 5% HPMC solution for 210 min, exhibited sustained release pattern within 8 h. Nifedipine release from MCC-based NSPs followed the Korsmeyer model and closely related to the microstructure of pellet. High stability of NSPs was confirmed after 6 months of accelerated stability test. Using commercially available sustained product as reference, bioequivalence study in beagle dogs was executed and two formulations were bioequivalent. This sustained release pellet formulation of nifedipine was advantageous with convenient and easy scaled-up preparation process.

A highly stable norcantharidin loaded lipid microspheres: preparation, biodistribution and targeting evaluation.

The purpose of this study was to prepare norcantharidin (NCTD)-loaded lipid microspheres (LMs) with a high encapsulation efficiency (EE) and stability during sterilization. The NCTD-phospholipid complex (NPC) was produced and characterized to increase the lipophilic properties of NCTD and a novel concentrated homogenization method was applied for the preparation of LMs. The results of the UV, DSC and IR investigations confirmed the formation of NPC. The oil-water partition coefficient (log P) of NPC was significantly increased with a value of -1.34 ± 0.06 at pH 7.4, nearly 224 times higher than that of NCTD. A concentrated emulsion was prepared based on a homogenization method and then diluted with water. After optimization of the NPC formation and emulsion preparation process, the EE was dramatically increased from 21.6% to 84.6%, and a highly sterilization stability was achieved with only a minor change in particle size from 168.2 ± 39.4 nm to 173.4 ± 43.5 nm. The tissue distribution of NPCLM was measured after intravenous administration to rats of a dose of 3.9 mg/kg with NCTD injection (NI) as the reference. Considerably increased concentrations of NCTD in the liver, spleen and lung were detected with NPCLM and the values were 1.67, 1.49 and 1.06 times higher than in the NI group, respectively while, in the kidney, the concentration was slightly reduced 0.96-fold. Overall, based on these techniques, this NPCLM with an improved EE and stability offers great promise in clinical applications and industrial-scale production along with a potentially increased targeting effect on the liver and reduced toxicity in the kidney.