Augmentation Inside the Network

12/19/2020
by   Maciej Sypetkowski, et al.
0

In this paper, we present augmentation inside the network, a method that simulates data augmentation techniques for computer vision problems on intermediate features of a convolutional neural network. We perform these transformations, changing the data flow through the network, and sharing common computations when it is possible. Our method allows us to obtain smoother speed-accuracy trade-off adjustment and achieves better results than using standard test-time augmentation (TTA) techniques. Additionally, our approach can improve model performance even further when coupled with test-time augmentation. We validate our method on the ImageNet-2012 and CIFAR-100 datasets for image classification. We propose a modification that is 30 than the flip test-time augmentation and achieves the same results for CIFAR-100.

READ FULL TEXT
POST COMMENT

Comments

There are no comments yet.

Authors

page 1

page 2

page 3

page 4

10/22/2020

Learning Loss for Test-Time Augmentation

Data augmentation has been actively studied for robust neural networks. ...
07/31/2017

Analysis and Optimization of Convolutional Neural Network Architectures

Convolutional Neural Networks (CNNs) dominate various computer vision ta...
04/08/2021

InAugment: Improving Classifiers via Internal Augmentation

Image augmentation techniques apply transformation functions such as rot...
02/24/2017

Changing Model Behavior at Test-Time Using Reinforcement Learning

Machine learning models are often used at test-time subject to constrain...
11/23/2020

When and Why Test-Time Augmentation Works

Test-time augmentation (TTA)—the aggregation of predictions across trans...
06/16/2021

ParticleAugment: Sampling-Based Data Augmentation

We present an automated data augmentation approach for image classificat...
06/14/2019

Fixing the train-test resolution discrepancy

Data-augmentation is key to the training of neural networks for image cl...
This week in AI

Get the week's most popular data science and artificial intelligence research sent straight to your inbox every Saturday.