HIGHLIGHTS
SUMMARY
Deep neural_networks, like convolutional neural_networks (CNNs), have been used in computer vision with numerous research applications, such as action recognition, object detection and localisation, face recognition, and image characterisation. The aim of these techniques is to reduce the complexity of a neural_network model during training, which is considered the main reason behind overfitting, especially when the model is trained on small datasets. For reviews on the deep learning approaches for data augmentation, see e_g_[31,32]. Techniques Model Task Findings Shijie et_al, 2017 CIFAR10; ImageNet (10 categories) GAN/WGAN, flipping, cropping, shifting, PCA jittering . . .
If you want to have access to all the content you need to log in!
Thanks :)
If you don't have an account, you can create one here.