Neural Architecture Search (NAS) has shown great potentials in automatically designing neural network architectures for real-time semantic segmentation. Unlike previous works that utilize a simplified search space with cell-sharing way, we introduce a new search space where a lightweight model can be more effectively searched by replacing the cell-sharing manner with cell-independent one. Based on this, the communication of local to global information is achieved through two well-designed modules. For local information exchange, a graph convolutional network (GCN) guided module is seamlessly integrated as a communication deliver between cells. For global information aggregation, we propose a novel dense-connected fusion module (cell) which aggregates long-range multi-level features in the network automatically. In addition, a latency-oriented constraint is endowed into the search process to balance the accuracy and latency. We name the proposed framework as Local-to-Global Information Communication Network Search (LGCNet). Extensive experiments on Cityscapes and CamVid datasets demonstrate that LGCNet achieves the new state-of-the-art trade-off between accuracy and speed. In particular, on Cityscapes dataset, LGCNet achieves the new best performance of 74.0\% mIoU with the speed of 115.2 FPS on Titan Xp.

翻译：神经架构搜索（NAS）在自动设计实时语义分割的神经网络架构方面展现出巨大潜力。与以往采用基于细胞共享方式的简化搜索空间的工作不同，我们引入了一种新的搜索空间，通过将细胞共享方式替换为细胞独立方式，可以更有效地搜索轻量级模型。在此基础上，通过两个精心设计的模块实现了局部到全局信息的通信。对于局部信息交换，我们无缝集成了一种图卷积网络（GCN）引导模块，作为细胞间的通信传递器。对于全局信息聚合，我们提出了一种新颖的密集连接融合模块（细胞），该模块自动聚合网络中的长距离多级特征。此外，我们在搜索过程中引入了延迟导向约束，以平衡精度和延迟。我们将所提出的框架命名为局部到全局信息通信网络搜索（LGCNet）。在Cityscapes和CamVid数据集上的大量实验表明，LGCNet在精度和速度之间实现了新的最先进平衡。特别是在Cityscapes数据集上，LGCNet在Titan Xp上以115.2 FPS的速度达到了74.0% mIoU的新的最佳性能。