Multi-objective Informative Frequency Band Selection Based on Negentropy-induced Grey Wolf Optimizer for Fault Diagnosis of Rolling Element Bearings.

Informative frequency band (IFB) selection is a challenging task in envelope analysis for the localized fault detection of rolling element bearings. In previous studies, it was often conducted with a single indicator, such as kurtosis, etc., to guide the automatic selection. However, in some cases, it is difficult for that to fully depict and balance the fault characters from impulsiveness and cyclostationarity of the repetitive transients. To solve this problem, a novel negentropy-induced multi-objective optimized wavelet filter is proposed in this paper. The wavelet parameters are determined by a grey wolf optimizer with two independent objective functions i.e., maximizing the negentropy of squared envelope and squared envelope spectrum to capture impulsiveness and cyclostationarity, respectively. Subsequently, the average negentropy is utilized in identifying the IFB from the obtained Pareto set, which are non-dominated by other solutions to balance the impulsive and cyclostationary features and eliminate the background noise. Two cases of real vibration signals with slight bearing faults are applied in order to evaluate the performance of the proposed methodology, and the results demonstrate its effectiveness over some fast and optimal filtering methods. In addition, its stability in tracking the IFB is also tested by a case of condition monitoring data sets.

DNA-functionalized metal or metal-containing nanoparticles for biological applications.

The conjugation of DNA molecules with metal or metal-containing nanoparticles (M/MC NPs) has resulted in a number of new hybrid materials, enabling a diverse range of novel biological applications in nanomaterial assembly, biosensor development, and drug/gene delivery. In such materials, the molecular recognition, gene therapeutic, and structure-directing functions of DNA molecules are coupled with M/MC NPs. In turn, the M/MC NPs have optical, catalytic, pore structure, or photodynamic/photothermal properties, which are beneficial for sensing, theranostic, and drug loading applications. This review focuses on the different DNA functionalization protocols available for M/MC NPs, including gold NPs, upconversion NPs, metal-organic frameworks, metal oxide NPs and quantum dots. The biological applications of DNA-functionalized M/MC NPs in the treatment or diagnosis of cancers are discussed in detail.

A Spatiotemporally Controlled Gene-Regulation Strategy for Combined Tumor Therapy Based on Upconversion Hybrid Nanosystem.

The lack of precise spatiotemporal gene modulation and therapy impedes progress in medical applications. Herein, a 980 nm near-infrared (NIR) light-controlled nanoplatform, namely URMT, is developed, which can allow spatiotemporally controlled photodynamic therapy and trigger the enzyme-activated gene expression regulation in tumors. URMT is constructed by engineering an enzyme-activatable antisense oligonucleotide, which combined with an upconversion nanoparticle (UCNP)-based photodynamic nanosystem, followed by the surface functionalization of triphenylphosphine (TPP), a mitochondria-targeting ligand. URMT allows for the 980 nm NIR light-activated generation of reactive oxygen species, which can induce the translocation of a DNA repair enzyme (namely apurinic/apyrimidinic endonuclease 1, APE1) from the nucleus to mitochondria. APE1 can recognize the basic apurinic/apyrimidinic (AP) sites in DNA double-strands and perform cleavage, thereby releasing the functional single-strands for gene regulation. Overall, an augmented antitumor effect is observed due to NIR light-controlled mitochondrial damage and enzyme-activated gene regulation. Altogether, the approach reported in this study offers high spatiotemporal precision and shows the potential to achieve precise and specific gene regulation for targeted tumor treatment.

Deep Hyperspherical Clustering for Skin Lesion Medical Image Segmentation.

Diagnosis of skin lesions based on imaging techniques remains a challenging task because data (knowledge) uncertainty may reduce accuracy and lead to imprecise results. This paper investigates a new deep hyperspherical clustering (DHC) method for skin lesion medical image segmentation by combining deep convolutional neural networks and the theory of belief functions (TBF). The proposed DHC aims to eliminate the dependence on labeled data, improve segmentation performance, and characterize the imprecision caused by data (knowledge) uncertainty. First, the SLIC superpixel algorithm is employed to group the image into multiple meaningful superpixels, aiming to maximize the use of context without destroying the boundary information. Second, an autoencoder network is designed to transform the superpixels' information into potential features. Third, a hypersphere loss is developed to train the autoencoder network. The loss is defined to map the input to a pair of hyperspheres so that the network can perceive tiny differences. Finally, the result is redistributed to characterize the imprecision caused by data (knowledge) uncertainty based on the TBF. The proposed DHC method can well characterize the imprecision between skin lesions and non-lesions, which is particularly important for the medical procedures. A series of experiments on four dermoscopic benchmark datasets demonstrate that the proposed DHC yields better segmentation performance, increasing the accuracy of the predictions while can perceive imprecise regions compared to other typical methods.