[1912.05027] SpineNet: Learning Scale-Permuted Backbone for Recognition and Localizationopen searchopen navigation menucontact arXivsubscribe to arXiv mailings

Convolutional neural networks typically encode an input image into a series of intermediate features with decreasing resolutions. While this structure is suited to classification tasks, it does not perform well for tasks requiring simultaneous recognition and localization (e.g., object detection). The encoder-decoder architectures are proposed to resolve this by applying a decoder network onto a backbone model designed for classification tasks. In this paper, we argue encoder-decoder architecture is ineffective in generating strong multi-scale features because of the scale-decreased backbone. We propose SpineNet, a backbone with scale-permuted intermediate features and cross-scale connections that is learned on an object detection task by Neural Architecture Search. Using similar building blocks, SpineNet models outperform ResNet-FPN models by ~3% AP at various scales while using 10-20% fewer FLOPs. In particular, SpineNet-190 achieves 52.5% AP with a MaskR-CNN detector and achieves 52

3 mentions: @YinCui1@windsor718@neuroecology
Date: 2020/07/01 03:51

Referring Tweets

@windsor718 従来の解像度をどんどん下げていくCNNをベースに、その各レイヤーを強化学習で並び替えて互いにつなぎ合わせたら計算効率も精度も上がったらしい。source: CVPR2020&Google AI blog t.co/0AKfYx1i0O
@YinCui1 Our recent work about Scale-Permuted Network (SpineNet) Paper (CVPR 2020): t.co/2CkKuWfU7D Code: t.co/JdXpOGNhEb t.co/24ZhWU6aAk

Related Entries

Read more [1905.11946] EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks
2 users, 33 mentions 2019/05/31 03:48
Read more [1910.00127] A Mobile Manipulation System for One-Shot Teaching of Complex Tasks in Homes
0 users, 2 mentions 2019/10/02 03:48
Read more CNN-generated images are surprisingly easy to spot... for now
0 users, 18 mentions 2020/03/02 12:51
Read more [2004.10934] YOLOv4: Optimal Speed and Accuracy of Object Detectionopen searchopen navigation menuco...
0 users, 26 mentions 2020/04/24 06:52
Read more [1909.11512] Synthetic Data for Deep Learningopen searchopen navigation menucontact arXivsubscribe t...
1 users, 10 mentions 2020/07/03 06:53