Fixing the train-test resolution discrepancy

Data-augmentation is key to the training of neural networks for image classification. This paper first shows that existing augmentations induce a significant discrepancy between the typical size of the objects seen by the classifier at train and test time... (read more)

PDF Abstract NeurIPS 2019 PDF NeurIPS 2019 Abstract

Results from the Paper


 Ranked #1 on Image Classification on iNaturalist (using extra training data)

     Get a GitHub badge
TASK DATASET MODEL METRIC NAME METRIC VALUE GLOBAL RANK USES EXTRA
TRAINING DATA
RESULT BENCHMARK
Fine-Grained Image Classification Birdsnap FixSENet-154 Accuracy 84.3% # 2
Fine-Grained Image Classification CUB-200-2011 FixSENet-154 Accuracy 88.7 # 4
Image Classification ImageNet FixPNASNet-5 Top 1 Accuracy 83.7% # 39
Top 5 Accuracy 96.8% # 20
Number of params 86.1M # 25
Image Classification ImageNet FixResNet-50 Top 1 Accuracy 79.1% # 99
Top 5 Accuracy 94.6% # 58
Number of params 25.6M # 64
Image Classification ImageNet FixResNet-50 CutMix Top 1 Accuracy 79.8% # 92
Top 5 Accuracy 94.9% # 49
Image Classification ImageNet FixResNeXt-101 32x48d Top 1 Accuracy 86.4% # 15
Top 5 Accuracy 98.0% # 7
Number of params 829M # 2
Image Classification ImageNet FixResNet-50 Billion Top 1 Accuracy 82.5% # 53
Top 5 Accuracy 96.6% # 23
Image Classification ImageNet ReaL FixResNeXt-101 32x48d Accuracy 89.73% # 7
Params 829M # 8
Image Classification iNaturalist FixSENet-154 Top 1 Accuracy 75.4% # 1
Fine-Grained Image Classification NABirds FixSENet-154 Accuracy 89.2% # 1
Fine-Grained Image Classification Oxford 102 Flowers FixInceptionResNet-V2 Top-1 Error Rate 4.3 # 3
Accuracy 95.7% # 12
Fine-Grained Image Classification Oxford-IIIT Pets FixSENet-154 Top-1 Error Rate 5.2 # 3
Accuracy 94.8% # 7
Fine-Grained Image Classification Stanford Cars FixSENet-154 Accuracy 94.4% # 14

Methods used in the Paper


METHOD TYPE
Average Pooling
Pooling Operations
ResNeXt Block
Skip Connection Blocks
Grouped Convolution
Convolutions
Bottleneck Residual Block
Skip Connection Blocks
Global Average Pooling
Pooling Operations
Residual Block
Skip Connection Blocks
Residual Connection
Skip Connections
ReLU
Activation Functions
Kaiming Initialization
Initialization
Max Pooling
Pooling Operations
1x1 Convolution
Convolutions
Convolution
Convolutions
Batch Normalization
Normalization
ColorJitter
Image Data Augmentation
Random Horizontal Flip
Image Data Augmentation
Random Resized Crop
Image Data Augmentation
FixRes
Image Scaling Strategies
ResNet
Convolutional Neural Networks
ResNeXt
Convolutional Neural Networks