NAS-FPN: Learning Scalable Feature Pyramid Architecture for Object Detection

被引：1420

作者：

Ghiasi, Golnaz ^{[1
]}

Lin, Tsung-Yi ^{[1
]}

Le, Quoc V. ^{[1
]}

机构：

[1] Google Brain, Mountain View, CA 94043 USA

来源：

2019 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR 2019) | 2019年

关键词：

D O I：

10.1109/CVPR.2019.00720

中图分类号：

TP18 [人工智能理论];

学科分类号：

140502 [人工智能];

摘要：

Current state-of-the-art convolutional architectures for object detection are manually designed. Here we aim to learn a better architecture of feature pyramid network for object detection. We adopt Neural Architecture Search and discover a new feature pyramid architecture in a novel scalable search space covering all cross-scale connections. The discovered architecture, named NAS-FPN, consists of a combination of top-down and bottom-up connections to fuse features across scales. NAS-FPN, combined with various backbone models in the RetinaNet framework, achieves better accuracy and latency tradeoff compared to state-of-the-art object detection models. NAS-FPN improves mobile detection accuracy by 2 AP compared to state-of-the-art SS-DLite with MobileNetV2 model in [32] and achieves 48.3 AP which surpasses Mask R-CNN [10] detection accuracy with less computation time.

引用

页码：7029 / 7038

页数：10