Point cloud analysis is very challenging, as the shape implied in irregular points is difficult to capture. In this paper, we propose RS-CNN, namely, Relation-Shape Convolutional Neural Network, which extends regular grid CNN to irregular configuration for point cloud analysis.The key to RS-CNN is learning from relation, i.e., the geometric topology constraint among points. Specifically, the convolutional weight for local point set is forced to learn a high-level relation expression from predefined geometric priors, between a sampled point from this point set and the others. In this way, an inductive local representation with explicit reasoning about the spatial layout of points can be obtained, which leads to much shape awareness and robustness. With this convolution as a basic operator, RS-CNN, a hierarchical architecture can be developed to achieve contextual shape-aware learning for point cloud analysis. Extensive experiments on challenging benchmarks across three tasks verify RS-CNN achieves the state of the arts.

Mn2+-doped ZnSe quantum dots (Mn:ZnSe d-dots) with a tunable photoluminescence (PL) peak position were made to be water soluble by coating them with a monolayer of mercaptopropionic acid, a very short hydrophilic thiol. If the dopant centers were located close to the surface, thiol-coating partially quenched the PL. With about 2-3 monolayers of pure ZnSe on the surface, the PL of d-dots was actually enhanced upon thiol coating. When the doping centers were placed reasonably inside a d-dot, with about four monolayers of pure ZnSe between the doping centers and the surface ligands, the thiol ligands did not quench the PL of the d-dots, even though they did completely quench the PL of intrinsic ZnSe quantum dots. The overall size of such d-dots/ligand complex is only about 7-8 nm, implying an excellent permeability in biological issues. These d-dots were found to be exceptionally stable against continuous UV radiation in air for at least 25 days. They were also stable in boiling water with air bubbling under room light for hours. Recognition of a biotin pattern by d-dots conjugated with avidine was carried to illustrate the suitability of these efficient (about 40% PL quantum yield), stable, small, and water-soluble d-dots as biomedical labeling reagents.

Point cloud processing is very challenging, as the diverse shapes formed by irregular points are often indistinguishable. A thorough grasp of the elusive shape requires sufficiently contextual semantic information, yet few works devote to this. Here we propose DensePoint, a general architecture to learn densely contextual representation for point cloud processing. Technically, it extends regular grid CNN to irregular point configuration by generalizing a convolution operator, which holds the permutation invariance of points, and achieves efficient inductive learning of local patterns. Architecturally, it finds inspiration from dense connection mode, to repeatedly aggregate multi-level and multi-scale semantics in a deep hierarchy. As a result, densely contextual information along with rich semantics, can be acquired by DensePoint in an organic manner, making it highly effective. Extensive experiments on challenging benchmarks across four tasks, as well as thorough model analysis, verify DensePoint achieves the state of the arts.

A dendron ligand with two carboxylate anchoring groups at its focal point and eight hydroxyl groups as its terminal groups was found to efficiently convert as-synthesized CdSe/CdS core-shell nanocrystals in toluene to water-soluble dendron-ligand stabilized nanocrystals (dendron nanocrystals). The resulting dendron nanocrystals retained 60% of the photoluminescence value of the original CdSe/CdS core-shell nanocrystals in toluene and were significantly brighter than the similar dendron nanocrystals with thiolate (deprotonated thiol group) as the anchoring group which retained just 10% of the photoluminescence value of the original CdSe/CdS core-shell nanocrystals in toluene. The carboxylate-based dendron nanocrystals survived UV irradiation in air for at least 13 days, about 9 times better than the thiolate-based dendron nanocrystals (35 h) and similar to that of the thiolate-based dendron-box stabilized CdSe/CdS core-shell nanocrystals (box nanocrystals). Upon UV irradiation, the dendron nanocrystals became even 2 times brighter than the original CdSe/CdS core-shell nanocrystals in toluene, and the UV-brightened PL can retain the brightness for at least several months. These stable and bright dendron nanocrystals were soluble in various aqueous media, including all common biological buffer solutions tested, for at least 1.5 years. In addition to their superior performance, the synthetic chemistry of carboxylate dendron ligands and the corresponding dendron nanocrystals is relatively simple and with high yield.