No abstract

No abstract

Semantic Scene Completion (SSC) refers to the task of inferring the 3D semantic segmentation of a scene while simultaneously completing the 3D shapes. We propose PALNet, a novel hybrid network for SSC based on single depth. PALNet utilizes a two-stream network to extract both 2D and 3D features from multi-stages using fine-grained depth information to efficiently captures the context, as well as the geometric cues of the scene. Current methods for SSC treat all parts of the scene equally causing unnecessary attention to the interior of objects. To address this problem, we propose Position Aware Loss(PA-Loss) which is position importance aware while training the network. Specifically, PA-Loss considers Local Geometric Anisotropy to determine the importance of different positions within the scene. It is beneficial for recovering key details like the boundaries of objects and the corners of the scene. Comprehensive experiments on two benchmark datasets demonstrate the effectiveness of the proposed method and its superior performance. Code and demo 1 are avaliable at: https://github.com/UniLauX/PALNet.

Many scientific and engineering applications involve finding more than one optimum. A comprehensive review of the existing works done in the field of multimodal function optimization was given and a critical analysis of the existing methods was also provided. Several techniques in solving multimodal function optimization problems were introduced, such as clearing, deterministic crowding, sharing, species conserving and so on. And we summarized defects of existing algorithms: lacking of self-adaptive adjustment function, requiring setting some parameters according to different problems, lacking of unified theoretical and experimental system to guide algorithms design and not maintaining the diversity of swarm. Moreover, most of existing multimodal particle swarm optimization algorithms which include SPSO, MSPSO, ESPSO, ANPSO, kPSO, MGPSO, AT-MGPSO, rpso, and SDD-PSO were described and compared and advantages and disadvantages existing in these algorithms were pointed out. Therefore, some ideas to improve the performance of multimodal function optimization algorithms were proposed

No abstract