We present DeepICP -a novel end-to-end learningbased 3D point cloud registration framework that achieves comparable registration accuracy to prior state-of-the-art geometric methods. Different from other keypoint based methods where a RANSAC procedure is usually needed, we implement the use of various deep neural network structures to establish an end-to-end trainable network. Our keypoint detector is trained through this end-to-end structure and enables the system to avoid the inference of dynamic objects, leverages the help of sufficiently salient features on stationary objects, and as a result, achieves high robustness. Rather than searching the corresponding points among existing points, the key contribution is that we innovatively generate them based on learned matching probabilities among a group of candidates, which can boost the registration accuracy. Our loss function incorporates both the local similarity and the global geometric constraints to ensure all above network designs can converge towards the right direction. We comprehensively validate the effectiveness of our approach using both the KITTI dataset and the Apollo-SouthBay dataset. Results demonstrate that our method achieves comparable or better performance than the state-of-the-art geometry-based methods. Detailed ablation and visualization analysis are included to further illustrate the behavior and insights of our network. The low registration error and high robustness of our method makes it attractive for substantial applications relying on the point cloud registration task.

Chronic nonhealing wounds have imposed serious challenges in the clinical practice, especially for the patients infected with multidrug-resistant microbes. Herein, we developed an ultrasmall copper sulfide (covellite) nanodots (CuS NDs) based dual functional nanosystem to cure multidrug-resistant bacteria-infected chronic nonhealing wound. The nanosystem could eradicate multidrug-resistant bacteria and expedite wound healing simultaneously owing to the photothermal effect and remote control of copper-ion release. The antibacterial results indicated that the combination treatment of photothermal CuS NDs with photothermal effect initiated a strong antibacterial effect for drug-resistant pathogens including methicillin-resistant Staphylococcus aureus (MRSA) and extended-spectrum β-lactamase Escherichia coli both in vitro and in vivo. Meanwhile, the released Cu 2+ could promote fibroblast cell migration and endothelial cell angiogenesis, thus accelerating wound-healing effects. In MRSA-infected diabetic mice model, the nanosystem exhibited synergistic wound healing effect of infectious wounds in vivo and demonstrated negligible toxicity and nonspecific damage to major organs. The combination of ultrasmall CuS NDs with photothermal therapy displayed enhanced therapeutic efficacy for chronic nonhealing wound in multidrug-resistant bacterial infections, which may represent a promising class of antibacterial strategy for clinical translation.

CRISPR-Cas9 gene editing has emerged as a powerful therapeutic technology, but the lack of safe and efficient in vivo delivery systems, especially for tissue-specific vectors, limits its broad clinical applications. Delivery of Cas9 ribonucleoprotein (RNP) owns competitive advantages over other options; however, the large size of RNPs exceeds the loading capacity of currently available delivery vectors. Here, we report a previously unidentified genome editing delivery system, named exosome RNP , in which Cas9 RNPs were loaded into purified exosomes isolated from hepatic stellate cells through electroporation. Exosome RNP facilitated effective cytosolic delivery of RNP in vitro while specifically accumulated in the liver tissue in vivo. Exosome RNP showed vigorous therapeutic potential in acute liver injury, chronic liver fibrosis, and hepatocellular carcinoma mouse models via targeting p53 up-regulated modulator of apoptosis ( PUMA ), cyclin E1 ( CcnE1 ), and K (lysine) acetyltransferase 5 ( KAT5 ), respectively. The developed exosome RNP provides a feasible platform for precise and tissue-specific gene therapies of liver diseases.

In Part I, a new theory for impact ionization that utilizes history-dependent ionization coefficients to account for the nonlocal nature of the ionization process has been described. In this paper, we will review this theory and extend it with the assumptions that are implicitly used in both the local-field theory in which the ionization coefficients are functions only of the local electric field and the new one. A systematic study of the noise characteristics of GaAs homojunction avalanche photodiodes with different multiplication layer thicknesses is also presented. It is demonstrated that there is a definite "size effect" for thin multiplication regions that is not well characterized by the local-field model. The new theory, on the other hand, provides very good fits to the measured gain and noise. The new ionization coefficient model has also been validated by Monte Carlo simulations.

No abstract