LIP: Local Importance-based Pooling

ICCV 2019  ·  Ziteng Gao, Li-Min Wang, Gangshan Wu ·

Spatial downsampling layers are favored in convolutional neural networks (CNNs) to downscale feature maps for larger receptive fields and less memory consumption. However, for discriminative tasks, there is a possibility that these layers lose the discriminative details due to improper pooling strategies, which could hinder the learning process and eventually result in suboptimal models... In this paper, we present a unified framework over the existing downsampling layers (e.g., average pooling, max pooling, and strided convolution) from a local importance view. In this framework, we analyze the issues of these widely-used pooling layers and figure out the criteria for designing an effective downsampling layer. According to this analysis, we propose a conceptually simple, general, and effective pooling layer based on local importance modeling, termed as {\em Local Importance-based Pooling} (LIP). LIP can automatically enhance discriminative features during the downsampling procedure by learning adaptive importance weights based on inputs. Experiment results show that LIP consistently yields notable gains with different depths and different architectures on ImageNet classification. In the challenging MS COCO dataset, detectors with our LIP-ResNets as backbones obtain a consistent improvement ($\ge 1.4\%$) over the vanilla ResNets, and especially achieve the current state-of-the-art performance in detecting small objects under the single-scale testing scheme. read more

PDF Abstract ICCV 2019 PDF ICCV 2019 Abstract

Datasets


Results from the Paper


Task Dataset Model Metric Name Metric Value Global Rank Result Benchmark
Object Detection COCO minival Faster R-CNN (LIP-ResNet-101) box AP 41.7 # 63
AP50 63.6 # 27
AP75 45.6 # 37
APS 25.2 # 35
APM 45.8 # 33
Object Detection COCO test-dev Faster R-CNN (LIP-ResNet-101-MD w FPN) box AP 43.9 # 91
AP50 65.7 # 58
AP75 48.1 # 79
APS 25.4 # 87
APM 46.7 # 86
APL 56.3 # 81
Image Classification ImageNet LIP-ResNet-101 Top 1 Accuracy 79.33% # 222
Top 5 Accuracy 94.6% # 109
Number of params 42.9M # 104
Image Classification ImageNet ResNet-50 (LIP Bottleneck-256) Top 1 Accuracy 78.15% # 269
Top 5 Accuracy 94.02% # 136
Number of params 25.8M # 132
Image Classification ImageNet LIP-DenseNet-BC-121 Top 1 Accuracy 76.64% # 294
Top 5 Accuracy 93.16% # 155
Number of params 8.7M # 176

Methods