Auto-DeepLab: Hierarchical Neural Architecture Search for Semantic Image Segmentation

TL;DR

Auto-DeepLab introduces a hierarchical neural architecture search method that optimizes both cell and network level structures, achieving state-of-the-art semantic segmentation results efficiently without ImageNet pretraining.

Contribution

The paper proposes a novel hierarchical NAS framework for semantic segmentation, including network level search space and an efficient gradient-based search formulation.

Findings

Achieves state-of-the-art results on Cityscapes, PASCAL VOC 2012, and ADE20K datasets.

Efficient search requiring only 3 P100 GPU days.

Outperforms previous methods without ImageNet pretraining.

Abstract

Recently, Neural Architecture Search (NAS) has successfully identified neural network architectures that exceed human designed ones on large-scale image classification. In this paper, we study NAS for semantic image segmentation. Existing works often focus on searching the repeatable cell structure, while hand-designing the outer network structure that controls the spatial resolution changes. This choice simplifies the search space, but becomes increasingly problematic for dense image prediction which exhibits a lot more network level architectural variations. Therefore, we propose to search the network level structure in addition to the cell level structure, which forms a hierarchical architecture search space. We present a network level search space that includes many popular designs, and develop a formulation that allows efficient gradient-based architecture search (3 P100 GPU days on Cityscapes images). We demonstrate the effectiveness of the proposed method on the challenging Cityscapes, PASCAL VOC 2012, and ADE20K datasets. Auto-DeepLab, our architecture searched specifically for semantic image segmentation, attains state-of-the-art performance without any ImageNet pretraining.