From source to target and back: symmetric bi-directional adaptive GAN

05/24/2017
by   Paolo Russo, et al.
0

The effectiveness of generative adversarial approaches in producing images according to a specific style or visual domain has recently opened new directions to solve the unsupervised domain adaptation problem. It has been shown that source labeled images can be modified to mimic target samples making it possible to train directly a classifier in the target domain, despite the original lack of annotated data. Inverse mappings from the target to the source domain have also been evaluated but only passing through adapted feature spaces, thus without new image generation. In this paper we propose to better exploit the potential of generative adversarial networks for adaptation by introducing a novel symmetric mapping among domains. We jointly optimize bi-directional image transformations combining them with target self-labeling. Moreover we define a new class consistency loss that aligns the generators in the two directions imposing to conserve the class identity of an image passing through both domain mappings. A detailed qualitative and quantitative analysis of the reconstructed images confirm the power of our approach. By integrating the two domain specific classifiers obtained with our bi-directional network we exceed previous state-of-the-art unsupervised adaptation results on four different benchmark datasets.

READ FULL TEXT

page 6

page 8

page 9

research
02/12/2020

Bi-Directional Generation for Unsupervised Domain Adaptation

Unsupervised domain adaptation facilitates the unlabeled target domain r...
research
12/10/2019

Learning Domain Adaptive Features with Unlabeled Domain Bridges

Conventional cross-domain image-to-image translation or unsupervised dom...
research
11/24/2017

CatGAN: Coupled Adversarial Transfer for Domain Generation

This paper introduces a Coupled adversarial transfer GAN (CatGAN), an ef...
research
12/13/2020

Bi-Classifier Determinacy Maximization for Unsupervised Domain Adaptation

Unsupervised domain adaptation challenges the problem of transferring kn...
research
09/25/2020

Deep Adversarial Transition Learning using Cross-Grafted Generative Stacks

Current deep domain adaptation methods used in computer vision have main...
research
06/21/2021

CUDA-GR: Controllable Unsupervised Domain Adaptation for Gaze Redirection

The aim of gaze redirection is to manipulate the gaze in an image to the...
research
03/25/2018

P2P-NET: Bidirectional Point Displacement Network for Shape Transform

We introduce P2P-NET, a general-purpose deep neural network which learns...

Please sign up or login with your details

Forgot password? Click here to reset