Current NAS-based semantic segmentation methods focus on accuracy improvements rather than light-weight design. In this paper, we proposed a two-stage framework to design our NAS-based RSPNet model for light-weight semantic segmentation. The first architecture search determines the inner cell structure, and the second architecture search considers exponentially growing paths to finalize the outer structure of the network. It was shown in the literature that the fusion of high- and low-resolution feature maps produces stronger representations. To find the expected macro structure without manual design, we adopt a new path-attention mechanism to efficiently search for suitable paths to fuse useful information for better segmentation. Our search for repeatable micro-structures from cells leads to a superior network architecture in semantic segmentation. In addition, we propose an RSP (recursive Stage Partial) architecture to search a light-weight design for NAS-based semantic segmentation. The proposed architecture is very efficient, simple, and effective that both the macro- and micro- structure searches can be completed in five days of computation on two V100 GPUs. The light-weight NAS architecture with only 1/4 parameter size of SoTA architectures can achieve SoTA performance on semantic segmentation on the Cityscapes dataset without using any backbones.
翻译:以NAS为基础的当前语义分解方法侧重于精确度改进,而不是轻量度设计。在本文中,我们提出了设计以NAS为基础的轻量语义分解的 RSPNet 模型的两阶段框架。第一个架构搜索决定了内部单元格结构,第二个架构搜索考虑了最终确定网络外部结构的指数增长路径。文献显示,高分辨率和低分辨率地貌图的混合产生更强的表示力。为了找到没有手工设计的预期宏观结构,我们采用了一个新的路径跟踪机制,以便高效率地寻找连接有用信息以更好地分解的合适路径。我们从细胞中重复的微型结构的搜索导致在语义分解中形成一个更高级的网络结构。此外,我们提出了一个RSP(分解阶段部分)结构,以搜索以最终确定网络外部结构。拟议的结构非常高效、简单和有效,即宏观和微观结构搜索可以在两个V100GPS的计算5天完成。我们从单元格中重复使用的微型和微型结构的微型微型结构,在不使用SO4级级结构中,只能使用SOTA级结构的光量级结构。