Sampling Strategies for GAN Synthetic Data

by   Binod Bhattarai, et al.

Generative Adversarial Networks (GANs) have been used widely to generate large volumes of synthetic data. This data is being utilized for augmenting with real examples in order to train deep Convolutional Neural Networks (CNNs). Studies have shown that the generated examples lack sufficient realism to train deep CNNs and are poor in diversity. Unlike previous studies of randomly augmenting the synthetic data with real data, we present our simple, effective and easy to implement synthetic data sampling methods to train deep CNNs more efficiently and accurately. To this end, we propose to maximally utilize the parameters learned during training of the GAN itself. These include discriminator's realism confidence score and the confidence on the target label of the synthetic data. In addition to this, we explore reinforcement learning (RL) to automatically search a subset of meaningful synthetic examples from a large pool of GAN synthetic data. We evaluate our method on two challenging face attribute classification data sets viz. AffectNet and CelebA. Our extensive experiments clearly demonstrate the need of sampling synthetic data before augmentation, which also improves the performance of one of the state-of-the-art deep CNNs in vitro.


page 1

page 4

page 7


G2R Bound: A Generalization Bound for Supervised Learning from GAN-Synthetic Data

Performing supervised learning from the data synthesized by using Genera...

GAN-based Synthetic Medical Image Augmentation for increased CNN Performance in Liver Lesion Classification

Deep learning methods, and in particular convolutional neural networks (...

AugLabel: Exploiting Word Representations to Augment Labels for Face Attribute Classification

Augmenting data in image space (eg. flipping, cropping etc) and activati...

How far generated data can impact Neural Networks performance?

The success of deep learning models depends on the size and quality of t...

Synthetic Data Generation for Fraud Detection using GANs

Detecting money laundering in gambling is becoming increasingly challeng...

SynCGAN: Using learnable class specific priors to generate synthetic data for improving classifier performance on cytological images

One of the most challenging aspects of medical image analysis is the lac...

Deep Adversarial Training for Multi-Organ Nuclei Segmentation in Histopathology Images

Nuclei segmentation is a fundamental task that is critical for various c...

Please sign up or login with your details

Forgot password? Click here to reset