Deep Depth Completion From Extremely Sparse Data: A Survey.

Depth completion aims at predicting dense pixel-wise depth from an extremely sparse map captured from a depth sensor, e.g., LiDARs. It plays an essential role in various applications such as autonomous driving, 3D reconstruction, augmented reality, and robot navigation. Recent successes on the task have been demonstrated and dominated by deep learning based solutions. In this article, for the first time, we provide a comprehensive literature review that helps readers better grasp the research trends and clearly understand the current advances. We investigate the related studies from the design aspects of network architectures, loss functions, benchmark datasets, and learning strategies with a proposal of a novel taxonomy that categorizes existing methods. Besides, we present a quantitative comparison of model performance on three widely used benchmarks, including indoor and outdoor datasets. Finally, we discuss the challenges of prior works and provide readers with some insights for future research directions.

Highly selective electrocatalytic reduction of nitrate to nitrogen in a chloride ion-free system by promoting kinetic mass transfer of intermediate products in a novel Pd-Cu adsorption confined cathode.

The mass transfer on the catalyst surface has a great influence on the selectivity of electrocatalytic nitrate reduction to nitrogen. In this study, a Pd-Cu adsorption confined nickel foam cathode is designed in the absence of both proton exchange membranes and chloride ions. The repulsion of the cathode enables intermediate products such as nitrite to accumulate in the confined region, resulting in an increase in the possibility of a second-order reaction to form nitrogen. The system can obtain more than 92% continuous N2 selectivity when it is used to treat 200 mg L-1 NO3--N under a current density of 8 mA cm-2, which is not only higher than those of semiconfined and nonconfined systems but also significantly better than the results obtained by Pd-Cu directly modified cathodes prepared by electrodeposition or impregnation. It is found that a high initial nitrate concentration and low current density are more beneficial for the accumulation of intermediates on Pd-Cu catalysts, thus improving the formation of nitrogen. A mechanism study reveals that the intermediates can completely occupy the active sites on the surface of Pd, avoiding the generation of active hydrogen, and therefore inhibiting the first-order reaction to produce ammonia. Moreover, the reducibility of Pd-Cu can also be gradually improved under the function of the cathode so that the system exhibits good stability. This study demonstrates an environmentally friendly and promising method for total nitrogen removal from industrial wastewater with high conductivity.

Neuron-derived neuropeptide Y fine-tunes the splenic immune responses.

The nervous and immune systems are closely entwined to maintain the immune balance in health and disease. Here, we showed that LPS can activate suprarenal and celiac ganglia (SrG-CG) neurons and upregulate NPY expression in rats. Single-cell sequencing analysis revealed that knockdown of the NPY gene in SrG-CG altered the proliferation and activation of splenic lymphocytes. In a neuron and splenocyte coculture system and in vivo experiments, neuronal NPY in SrG-CG attenuated the splenic immune response. Notably, we demonstrated that neuronal NPF in Drosophila exerted a conservative immunomodulatory effect. Moreover, numerous SNPs in NPY and its receptors were significantly associated with human autoimmune diseases, which was further supported by the autoimmune disease patients and mouse model experiments. Together, we demonstrated that NPY is an ancient language for nervous-immune system crosstalk and might be utilized to alleviate inflammatory storms during infection and to modulate immune balance in autoimmune diseases.