Geometry Sharing Network for 3D Point Cloud Classification and Segmentation

In spite of the recent progresses on classifying 3D point cloud with deep CNNs, large geometric transformations like rotation and translation remain challenging problem and harm the final classification performance. To address this challenge, we propose Geometry Sharing Network (GS-Net) which effectively learns point descriptors with holistic context to enhance the robustness to geometric transformations. Compared with previous 3D point CNNs which perform convolution on nearby points, GS-Net can aggregate point features in a more global way. Specially, GS-Net consists of Geometry Similarity Connection (GSC) modules which exploit Eigen-Graph to group distant points with similar and relevant geometric information, and aggregate features from nearest neighbors in both Euclidean space and Eigenvalue space. This design allows GS-Net to efficiently capture both local and holistic geometric features such as symmetry, curvature, convexity and connectivity. Theoretically, we show the nearest neighbors of each point in Eigenvalue space are invariant to rotation and translation. We conduct extensive experiments on public datasets, ModelNet40, ShapeNet Part. Experiments demonstrate that GS-Net achieves the state-of-the-art performances on major datasets, 93.3% on ModelNet40, and are more robust to geometric transformations.

Research on Simulation Technique of Shoal Protection Works in Middle and Lower Reaches of Yangtze River

The shoal protection works in the middle and lower Yangtze River is divided three types, i.e. riprap in bulk, dam body and soft mattress. As a new kind of shoal protection works, because the application time is very short, therefore the beach protection effect and the improvement measure have not thorough researched. In order to better apply to channel regulation project, the simulation technique of the X-type flexible mattress is thorough researched, the aspects of geometry similarity, gravity similarity, plan layout similarity and river bed deformation similarity are solved, the physical simulation technology problem is realized.

Research on the Cooperative CAPP Modularity Technique Supporting the ETO Product

The ETO product’s modularity manufacture needs the ETO product’s modular design as a foundation. By fully distinguishing and exploiting the geometry similarity, the structure similarity, the function similarity and the process similarity of the multi-ETO products, using the methods of standardization, modularization and serialization, the ETO products as well as the corresponding manufacture systems are modularized. In view of various standard components and parts produced by the modular design and manufacture, the products needed for the customers are directly assembled using a set of the advanced operation flow suitable for the modular assembly, and selecting the appropriate standard parts. Thus the production cost, the assembly cost as well as the inventory cost and so on are reduced. The modular manufacture is using the modular technology in the manufacture process to form the overall integrated production and assembly system of the components, and to recompose the manufacture process promptly according to the specific production. This paper proposed the optimization method of production organization, established the CAPP system’s architecture based on the modularity technique supporting the ETO product and developed the cooperative CAPP platform of the multi-functional hydraulic lift machinery.