Fast Neural Architecture Search of Compact Semantic Segmentation Models via Auxiliary Cells

CVPR 2019 Vladimir NekrasovHao ChenChunhua ShenIan Reid

Automated design of neural network architectures tailored for a specific task is an extremely promising, albeit inherently difficult, avenue to explore. While most results in this domain have been achieved on image classification and language modelling problems, here we concentrate on dense per-pixel tasks, in particular, semantic image segmentation using fully convolutional networks... (read more)

PDF Abstract CVPR 2019 PDF CVPR 2019 Abstract
TASK DATASET MODEL METRIC NAME METRIC VALUE GLOBAL RANK RESULT BENCHMARK
Monocular Depth Estimation NYU-Depth V2 FastDenseNas-arch1 RMSE 0.526 # 10
Monocular Depth Estimation NYU-Depth V2 FastDenseNas-arch2 RMSE 0.525 # 9
Monocular Depth Estimation NYU-Depth V2 FastDenseNas-arch0 RMSE 0.523 # 8
Semantic Segmentation PASCAL VOC 2012 val FastDenseNas-arch0 mIoU 78.0% # 9
Semantic Segmentation PASCAL VOC 2012 val FastDenseNas-arch1 mIoU 77.1% # 12
Semantic Segmentation PASCAL VOC 2012 val FastDenseNas-arch2 mIoU 77.3% # 11

Methods used in the Paper