STGAN: A Unified Selective Transfer Network for Arbitrary Image Attribute Editing

By Ming Liu, Yukang Ding, Min Xia, Xiao Liu, Errui Ding, Wangmeng Zuo, Shilei Wen. Computer Vision and Pattern Recognition, 2019

366 citations recorded by the source corpus.

Arbitrary attribute editing generally can be tackled by incorporating encoder-decoder and generative adversarial networks. However, the bottleneck layer in encoder-decoder usually gives rise to blurry and low quality editing result. And adding skip connections improves image quality at the cost of weakened attribute manipulation ability. Moreover, existing methods exploit target attribute vector to guide the flexible translation to desired target domain. In this work, we suggest to address these issues from selective transfer perspective. Considering that specific editing task is certainly only related to the changed attributes instead of all target attributes, our model selectively takes the difference between target and source attribute vectors as input. Furthermore, selective transfer units are incorporated with encoder-decoder to adaptively select and modify encoder feature for enhanced attribute editing. Experiments show that our method (i.e., STGAN) simultaneously improves attribute manipulation accuracy as well as perception quality, and performs favorably against state-of-the-arts in arbitrary face attribute editing and season translation.

Subfields: Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/4ecb1adad81d4ad7cac8908f464d21b2f441e868

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

Citations in the corpus

Generative Adversarial Networks for Face Generation: A Survey (2022, background): “STGAN [85] incorporates the difference attribute vector and selective transfer units (STUs) in an encoder-decoder network.”
Towards Flexible Blind JPEG Artifacts Removal (2021, background): “text-to-image generation [26, 35, 43] and facial attribute editing [9, 21, 27].”
Trustworthy AI: From Principles to Practices (2021, background): “Similar ideas have been used in generative networks to gain control over attributes, such as the gender, age, and ethnicity, in a face generator [221].”
Collaborative Learning for Faster StyleGAN Embedding (2020, methodology): “Generative Adversarial Networks [12] has been widely applied in various image processing tasks to synthesize realistic images, such as image-to-image [15, 42, 42, 37, 25], and semantic attribute editting [8, 23, 33, 30, 24].”
Image-to-Image Translation: Methods and Applications (2021, background): “style transfer artistic:[113], [1] artistic:[15] artistic:[2], [36], [136], [137], [145]; photo-realistic: [102], [139], [2], [75], [136], [137], artistic:[152]; photo-realistic: [90], [76], [77], [77], [143], [152], [154] \ artistic:[155], [81], [166]; photo-realistic: [155], [156], [81] artistic:[”
DeepFakes and Beyond: A Survey of Face Manipulation and Fake Detection (2020, background): “Other recent GAN approaches that improve both the image quality of the fake images and the control editing of the parameters are InterFaceGAN [40], UGAN [103], STGAN [92], and AttGAN [91].”
Frequency-Aware Deepfake Detection: Improving Generalizability through Frequency Space Domain Learning (2024, background): “AttGAN BEGAN CramerGAN InfoMaxGAN MMDGAN RelGAN S3GAN SNGAN STGAN Mean Acc.”
Learning Self-Consistency for Deepfake Detection (2020, methodology): “It can be done by either using entire face synthesis techniques [17, 29, 4] that directly output the whole fake image, such as GAN, or future development of stitching methods that completely removes or ambiguates the source feature.”
Disentangled Face Attribute Editing via Instance-Aware Latent Space Search (2021, methodology): “Here, we provide more qualitative results compared to the state-of-the-art methods, including the semantic direction based methods (InterfaceGAN [ Shen et al. , 2020 ] and GANSpace [ H ¨ ark ¨ onen et al. , 2020 ] ), and the conditional-GAN based method [ Liu et al. , 2019 ] on GAN-generated images.”
A Review on Generative Adversarial Networks: Algorithms, Theory, and Applications (2020, background): “Other related works can be found in [246], [247].”
Deepfakes generation and detection: state-of-the-art, open challenges, countermeasures, and way forward (2021, methodology): “[164] proposed the STGAN model that incorporated an attribute difference indicator and a selective transfer unit with an encoder-decoder to adaptively select and modify the encoded features.”