AttGAN: Facial Attribute Editing by Only Changing What You Want
The DOI matches its Crossref record; 47 of the 52 references it cites were found in a registry or an index.
Abstract
Facial attribute editing aims to manipulate single or multiple attributes on a given face image, i.e., to generate a new face image with desired attributes while preserving other details. Recently, the generative adversarial net (GAN) and encoder-decoder architecture are usually incorporated to handle this task with promising results. Based on the encoder-decoder architecture, facial attribute editing is achieved by decoding the latent representation of a given face conditioned on the desired attributes. Some existing methods attempt to establish an attribute-independent latent representation for further attribute editing. However, such attribute-independent constraint on the latent representation is excessive because it restricts the capacity of the latent representation and may result in information loss, leading to over-smooth or distorted generation. Instead of imposing constraints on the latent representation, in this work, we propose to apply an attribute classification constraint to the generated image to just guarantee the correct change of desired attributes, i.e., to change what you want. Meanwhile, the reconstruction learning is introduced to preserve attribute-excluding details, in other words, to only change what you want. Besides, the adversarial learning is employed for visually realistic editing. These three components cooperate with each other forming an effective framework for high quality facial attribute editing, referred as AttGAN. Furthermore, the proposed method is extended for attribute style manipulation in an unsupervised manner. Experiments on two wild datasets, CelebA and LFW, show that the proposed method outperforms the state-of-the-art on realistic attribute editing with other facial details well preserved.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The check wasn’t confident either way, so it gives no verdict — never disputed.
In the citing paper
“Other meth- ods such as lip-syncing approaches [67–69] and im- age quality metrics with support vector machine (SVM) [70] produce very high error rate when applied to detect deepfake videos from this newly produced dataset.”
From Deep learning for deepfakes creation and detection: A survey (2022) · cites this paper as [70]
Experimental — accuracy not yet measuredIn the citing paper
“AttGAN [41] is a technology for facial attribute manipulation; its attribute awareness enables precise and high-quality attribute changes, making it valuable for applications like face-swapping and age progression or regression.”
From Deepfake Attacks: Generation, Detection, Datasets, Challenges, and Research Directions (2023) · cites this paper as [41]
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
U-Net: Convolutional Networks for Biomedical Image Segmentation
Ronneberger, O., Fischer, P., & Brox, T. (2015) · Lecture notes in computer science · doi:10.1007/978-3-319-24574-4_28 · Cited by 99,404
Adam: A Method for Stochastic Optimization
Kingma, D. P., & Ba, J. (2014) · arXiv · doi:10.48550/arxiv.1412.6980 · Cited by 98,946 · cited 1×
“The model is trained by Adam optimizer [41] (β1 = 0.”
MethodBatch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift
Ioffe, S., & Szegedy, C. (2015) · arXiv · doi:10.48550/arxiv.1502.03167 · Cited by 27,818
Unpaired image-to-image translation using cycle-consistent adversarial networks
Zhu, J.-Y., Park, T., Isola, P., & Efros, A. A. (2017) · 2017 IEEE International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2017.244 · Cited by 23,602 · cited 6×
“We employ AttGAN on a season dataset [37] and a painting dataset [21] and the results are shown in Fig.”
BackgroundMethodInfluentialImage-to-Image Translation with Conditional Adversarial Networks
Isola, P., Zhu, J.-Y., Zhou, T., & Efros, A. A. (2017) · 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2017.632 · Cited by 22,930 · cited 1×
“We also employ the U-Net [38] like symmetric skip connections between the encoder and decoder, which have been shown to produce high quality results on the image translation task [39].”
MethodAuto-Encoding Variational Bayes
Kingma, D. P., & Welling, M. (2013) · arXiv · doi:10.48550/arxiv.1312.6114 · Cited by 20,067 · cited 2×
“In VAE/GAN [7], GAN [6] and VAE [5] are combined to learn a latent representation and a decoder.”
BackgroundLabaca-Castro, R. (2023) · Machine Learning under Malware Attack · doi:10.1007/978-3-658-40442-0_9 · Cited by 19,272
GAN(Generative Adversarial Nets)
(2017) · Journal of Japan Society for Fuzzy Theory and Intelligent Informatics · doi:10.3156/jsoft.29.5_177_2 · Cited by 19,270
FaceNet: A unified embedding for face recognition and clustering
Schroff, F., Kalenichenko, D., & Philbin, J. (2015) · 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2015.7298682 · Cited by 15,267 · cited 1×
“For conventional face recognition [1], [2] and facial attribute prediction [3], [4] tasks, significant advances have been made along with the development of deep convolutional neural networks (CNNs) and large scale labeled datasets.”
BackgroundUnsupervised representation learning with deep convolutional generative adversarial networks
Radford, A., Metz, L., & Chintala, S. (2015) · arXiv · doi:10.48550/arxiv.1511.06434 · Cited by 12,235 · cited 2×
“DCGAN [27] uses CNN and batch normalization [28] for stable training.”
Method
Cited by
Items 601–700 of 848
Testing the Channels of Convolutional Neural Networks
Choi, K. Y., Son, D., Kim, Y., & Seo, J. (2023) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v37i12.26726 · Cited by 2
Cascaded face super‐resolution with shape and identity priors
Zeng, D., Li, Z., Xiao, Y., Jiang, W., Wang, X., Liu, J., & Tang, B. (2023) · IET Image Processing · doi:10.1049/ipr2.12863 · Cited by 2
Data Augmentation Techniques for Facial Image Generation: A Brief Literature Review
Salais, B. E. C., Florencia, R., García, V., & Sánchez-Solís, J. P. (2023) · Studies in big data · doi:10.1007/978-3-031-38325-0_9 · Cited by 2
ERAT:Eyeglasses removal with attention
Zhang, H., & Guo, J. (2024) · Pattern Recognition · doi:10.1016/j.patcog.2024.110970 · Cited by 2
Pixel Sampling for Style Preserving Face Pose Editing
Yin, X., Huang, D., Yang, H., Fu, Z., Wang, Y., & Chen, L. (2020) · IEEE International Joint Conference on Biometrics (IJCB) · doi:10.1109/ijcb48548.2020.9304867 · Cited by 2
“In multi-domain image translation tasks, the domain information is provided either to the bottleneck layer of the auto-encoder [48, 16, 39], or to the entry of the encoder/generator [6, 33, 17], by simply concatenating the domain label with the features or input images.”
SATGAN: Augmenting Age Biased Dataset for Cross-Age Face Recognition
Liu, W., Chen, W., Zhu, Y., & Shen, L. (2021) · Proceedings - International Conference on Pattern Recognition/Proceedings/International Conference on Pattern Recognition · doi:10.1109/icpr48806.2021.9412084 · Cited by 2
“StarGAN [16] and AttGAN [17] encoded the image into feature map and utilized a cycle-consistent loss to achieve a good result in cross domain translation.”
MethodTang, S., Zhou, X., He, X., & Ma, Y. (2021) · Proceedings - International Conference on Pattern Recognition/Proceedings/International Conference on Pattern Recognition · doi:10.1109/icpr48806.2021.9411925 · Cited by 2
“There are some other approaches that make use of an auxiliary classifier to make the generator utilize the categorical labels [15], [16], [17] or adopt task-specific techniques [18].”
BackgroundMethodEyeglass Remover Network based on a Synthetic Image Dataset
Kang, S.-J., & Hahn, T. (2021) · KSII Transactions on Internet and Information Systems · doi:10.3837/tiis.2021.04.016 · Cited by 3
“He et al. proposed AttGAN [17], which improved the attribute editing performance by employing only the necessary adversarial loss, attribute classification loss, and reconstruction loss.”
BackgroundMethodReview on deep learning based techniques for person re-identification
Parkhi, A. A., & Khobragade, A. S. (2022) · 3C TIC Cuadernos de desarrollo aplicados a las TIC · doi:10.17993/3ctic.2022.112.208-223 · Cited by 3
Masked Based Unsupervised Content TransferPreprint
Mokady, R., Benaim, S., Wolf, L., & Bermano, A. H. (2020) · arXiv (Cornell University) · Cited by 2
Generative Transition Mechanism to Image-to-Image Translation via Encoded TransformationPreprint
Shi, Y., Zhou, X., Liu, P., & Tsang, I. W.-H. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2103.05193 · Cited by 3
“To achieve this goal, previous works either adopt corresponding attribute classifiers to control the attributes of translated results [10, 3, 28, 1, 19, 8, 13], or design specific loss function, e.”
BackgroundMethodInfluentialWhy Adopting Regularization and Normalization For Generative Adversarial Networks: A SurveyPreprint
Li, Z., Wu, X., Tao, R., Xia, P., Chen, H., & Li, B. (2020) · arXiv (Cornell University) · Cited by 2
Yu, C., & Wang, W. (2021) · arXiv (Cornell University) · Cited by 2
Sun, J., Bhattarai, B., Chen, Z., & Kim, T. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2111.09298 · Cited by 3
“We split the dataset into 28,000 for training, 500 for validation, and 1,500 for testing, again following [9].”
BackgroundMethodResultInfluentialFactored couplings in multi-marginal optimal transport via difference of convex programmingPreprint
Tran, Q. H., Janati, H., Redko, I., Flamary, R., & Courty, N. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2110.00629 · Cited by 2
“…far less popular than the classic OT with two marginals, MMOT is a very useful framework on its own with some notable recent applications in generative adversarial networks [Cao et al., 2019], clustering [Mi and Bento, 2020] and domain adaptation [Hui et al., 2018, He et al., 2019], to name a few.”
Deep neural network architectures for user-centered design concept generation and evaluation
Yuan, C. (2022) · Digital Repository Service (Northeastern University) · doi:10.17760/d20449078 · Cited by 3
Vidanaarachchi, R., Seneviratne, S., Kasthuriarachchi, N., Wijnands, J. S., & Rasnayaka, S. (2021) · International Conference on Electrical, Computer and Energy Technologies (ICECET) · doi:10.1109/icecet52533.2021.9698503 · Cited by 1
Content-Aware Latent Semantic Direction Fusion for Multi-Attribute Editing
Wei, X., Tang, Y., & Wu, S. (2023) · IEEE Multimedia · doi:10.1109/mmul.2023.3285550 · Cited by 1
Fan, Z., Zhang, Z., Yang, S., Zhong, C., Cao, M., & Xia, S. (2023) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51070.2023.01904 · Cited by 1
“For example, pix2pix [32] learns image-to-image translation with a CGAN architecture in a supervisory manner with paired data; Attgan [29] is able to change the specified attribute of a facial image.”
MethodZhu, S., Liang, J., & Liu, H. (2024) · Computer Animation and Virtual Worlds · doi:10.1002/cav.2273 · Cited by 1
Inversion-Based Face Swapping With Diffusion Model
Yoo, D., Lee, H.-C., & Kim, J. (2024) · IEEE Access · doi:10.1109/access.2024.3519163 · Cited by 6
Navigating h-Space for Multi-Attribute Editing in Diffusion Models
Park, J. M., Shaheryar, M., Lee, S., & Jung, S. K. (2025) · International Conference on Artificial Intelligence in Information and Communication (ICAIIC) · doi:10.1109/icaiic64266.2025.10920714 · Cited by 2
“Generative adversarial networks (GANs) [2] have demonstrated significant potential in image synthesis and semantic editing, leveraging latent space representations to achieve fine-grained attribute control [3]–[5].”
Liu, Y., Ouyang, X., & Cui, X. (2025) · Displays · doi:10.1016/j.displa.2025.103062 · Cited by 2
Synthetic Realities: A Critical Review of Deepfake Techniques
Yadav, M., Sheoran, S. K., & Singh, G. (2025) · International Conference on Cognitive Computing in Engineering, Communications, Sciences and Biomedical Health Informatics (IC3ECSBHI) · doi:10.1109/ic3ecsbhi63591.2025.10991244 · Cited by 1
“Alternative methods for generating deepfakes include DCGAN, GHOST [7], and AttGAN [8].”
Cao, Z., Wang, Z., Wang, R., Yang, Y., Tian, F., Wu, G., & Suzuki, A. (2025) · doi:10.1145/3746265.3759661 · Cited by 2
“Target models : In our training on GAN-based forgery models, we follow the configuration in CMUA [15], utilizing four models: HiSD [24], AttGAN [10], StarGAN [4], and AGGAN [36] to obtain universal perturbation.”
Compression-Resistant Adversarial Perturbation for Real-World Proactive Defense Against Deepfakes
Feng, Y., & Huang, F. (2025) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2025.3626505 · Cited by 1
“2) Deepfake Models: Five representative Deepfake models, namely StarGAN [30], AttentionGAN [43], AttGAN [31], Fixed-PointFAN [32], and HiSD [33] are chosen for our experiments.”
A Review of Deepfake Technology in Physical Health Management and Application
Fan, T., & Moghimi, M. M. (2026) · International Journal of Intelligent Systems · doi:10.1155/int/9983200 · Cited by 1
An analysis of deepfake technology: A systematic review of creation, detection, and mitigation
Singh, A., & Jaiswal, A. K. (2026) · Applied Soft Computing · doi:10.1016/j.asoc.2026.115644 · Cited by 2
CMUA++: A cross-model universal active defense framework for combating deepfake
Ye, X., Huang, H., Zhou, X., Wang, Y., & Ma, K. (2026) · Pattern Recognition · doi:10.1016/j.patcog.2026.114364 · Cited by 1
Multi-attributed Face Synthesis for One-Shot Deep Face Recognition
Shaheryar, M., Laishram, L., Lee, J. T., & Jung, S. K. (2023) · Communications in computer and information science · doi:10.1007/978-981-99-4914-4_1 · Cited by 1
StyleAU: StyleGAN based Facial Action Unit Manipulation for Expression Editing
Guo, Y., Hou, X., Liu, F., Shen, L., Wang, L., Wang, Z., & Liu, P. (2023) · IEEE International Joint Conference on Biometrics (IJCB) · doi:10.1109/ijcb57857.2023.10448612 · Cited by 3
“Recently, several works [4, 10, 34, 26, 22, 5] have achieved facial expression editing based on Generative Ad-versarial Nets (GANs).”
BackgroundGenerative Transferable Universal Adversarial Perturbation for Combating Deepfakes
Guo, Y., Wang, X., Fu, X., Liu, J., Li, Z., Chai, Y., & Han, J. (2024) · International Conference on Computer Supported Cooperative Work in Design (CSCWD) · doi:10.1109/cscwd61410.2024.10580713 · Cited by 5
“AttGAN [7] introduces an attribute classification constraint to ensure accurate transformation of specific attributes during image editing while keeping others unchanged.”
BackgroundEvasion on general GAN-generated image detection by disentangled representation
Chan, P. P. K., Zhang, C., Chen, H., Deng, J., Meng, X. J., & Yeung, D. S. (2024) · Information Sciences · doi:10.1016/j.ins.2024.121267 · Cited by 4
HyperplaneGAN: a unified consistent translation framework for facial attribute editing
Li, D., Deng, H., Qin, P., Chen, W., & Feng, G. (2024) · Multimedia Tools and Applications · doi:10.1007/s11042-024-20062-9 · Cited by 1
“…transfer model that is designed to transfer attributes along with corresponding styles between exemplars and target images, StarGAN [17], AttGAN [18] and STGAN [19] are semantic-level attribute transfer models that manipulate different attributes of target images by using different target…”
InfluentialSub-r-pa, C., Chen, R.-C., & Fan, M.-Z. (2024) · PeerJ Computer Science · doi:10.7717/peerj-cs.2438 · Cited by 1
Generation With Nuanced Changes: Continuous Image-to-Image Translation With Adversarial Preferences
Yao, Y., Pan, Y., Tsang, I. W.-H., & Yao, X. (2024) · IEEE Transactions on Artificial Intelligence · doi:10.1109/tai.2024.3497915 · Cited by 1
“[13], generative adversarial networks (GANs) [8], [9], [14], [15], [21], diffusion models [22].”
InfluentialImproving Sequential DeepFake Detection with Local information enhancement
Dong, L., Xu, Y., Zhong, J., Qi, Z., & Zhang, W. (2024) · ACM International Conference on Multimedia in Asia (MMAsia) · doi:10.1145/3696409.3700276 · Cited by 1
“To address this issue, researchers add skip connections between the encoder and decoder [8], [11].”
Controllable face soft-biometric privacy enhancement based on attribute disentanglement
Huang, W., Yao, Z., Jin, B., Chen, Z., & Wang, Y. (2025) · The Journal of Supercomputing · doi:10.1007/s11227-025-07134-9 · Cited by 1
“StarGAN v2 [23] adopts an autoencoder to generate target attribute faces based on input images and target labels, while AttGAN [24] extends attribute manipulation by not only adding or removing glasses but also generating various styles.”
HAIFAI: Human-AI Interaction for Mental Face Reconstruction
Strohm, F., Bâce, M., & Bulling, A. (2025) · ACM Transactions on Interactive Intelligent Systems · doi:10.1145/3725891 · Cited by 2
“Although less invasive and costly, this first method required prior knowledge of the target face, which is prohibitive for most practical applications.”
Three-in-One: Robust Enhanced Universal Transferable Anti-Facial Retrieval in Online Social Networks
Lv, Y., Tang, L., Ye, D., Xie, C., Deng, J., He, Y., & Shen, S. (2025) · IEEE Transactions on Information Forensics and Security · doi:10.1109/tifs.2025.3557737 · Cited by 1
“For attribute editing, we use AttGAN [49] to test the impact of TOAP on facial attribute editing and record the di ff erences between original generated images and adversarial generated images, measured by L 2 and L 2 mask , where smaller values indicate smaller di ff erences between the generated…”
Causality-driven attribute editing with bidirectional generative network
You, D., Zhao, Y. P., Shen, J., Li, Z., Chen, Z., & Wu, X. (2025) · Expert Systems with Applications · doi:10.1016/j.eswa.2025.127965 · Cited by 1
Revisiting DIRE: towards universal AI-generated image detection
Lin, H., Qin, J., Wu, X., Chen, T., & Yang, Z. (2025) · Neural Networks · doi:10.1016/j.neunet.2025.108084 · Cited by 6
Counterfactual generation with joint causal representation for generative adversarial networks
You, D.-L., Lu, C., Wu, Z.-J., Ge, X., & Wu, D. (2025) · Neural Networks · doi:10.1016/j.neunet.2025.108156 · Cited by 2
LoRA Patching: Exposing the Fragility of Proactive Defenses Against Deepfakes
Qu, Z., Guo, Y., Hu, Q., & Lu, W. (2025) · IEEE Signal Processing Letters · doi:10.1109/lsp.2025.3643352 · Cited by 1
“2) Baselines: We select StarGAN [28], AttGAN [29], and HiSD [30] as target Deepfake models, and employ four proactive defenses, including Disrupting [9], PG [10], CMUA [11], DF-RAP [13], RAW [14], and ID-Guard [15] to generate adversarial perturbations.”
Defending Deepfake via Texture Feature Perturbation
Zhang, X., Chen, C., & Wang, T. (2025) · Conference proceedings/Conference proceedings - IEEE International Conference on Systems, Man, and Cybernetics · doi:10.1109/smc58881.2025.11343336 · Cited by 1
“AttentionGAN [21] enhances local detail editing by using spatial attention maps to isolate target regions, while AttGAN [22] disentangles attribute and identity features, enforcing target attributes via classification loss.”
Liu, B., Cui, X., Zeng, R., Duan, W., Liu, C., Qian, J., Tang, L., & Gan, H. (2026) · Journal of King Saud University - Computer and Information Sciences · doi:10.1007/s44443-026-00575-z · Cited by 1
“Existing GAN-based methods are typically categorized into two main types: image space-based methods, which work directly in image space, and latent space-based methods, which operate in a learned latent space.”
Unsupervised Domain Adaptation-Based Cross-Type Deepfake Image Detection
Wang, Q., Wang, X., Liu, Z., Bai, N., Zhao, M., & Pang, S. (2026) · IEEE Transactions on Image Processing · doi:10.1109/tip.2026.3685445 · Cited by 3
“(2) Attribute manipulation datasets: The four datasets for attribute manipulation type are AttGAN [67], FaceAPP [68], StarGAN1 [69], and StarGAN2 [70].”
FTGID: Fine-Grained Text-Driven Framework for Universal Generative Image Detection
Huang, Z., Lin, L., Chen, B., Wang, Y., & Zhao, T. (2026) · IEEE Transactions on Image Processing · doi:10.1109/tip.2026.3685840 · Cited by 1
“In addition, we construct the GANGen dataset, which includes nine GAN-based face generation methods (AttGAN [53], BEGAN [54], CramerGAN [55], InfoMaxGAN [56], MMDGAN [57], Rel-GAN [58], S3GAN [59], SNGAN [60], STGAN [61]), to further evaluate adaptability across diverse generative scenarios.”
Wavelet transform-based versatile watermarking for facial manipulation source tracing and detection
Zhang, Y., Lin, W., Xu, J., Shi, L., Xu, W., Xu, Y., & Kou, F. (2026) · Information Processing & Management · doi:10.1016/j.ipm.2026.104852 · Cited by 1
Onyekwere, C. G.-L., Otuu, O. O., & Githinji, S. M. (2026) · World Journal of Advanced Research and Reviews · doi:10.30574/wjarr.2026.30.2.1387 · Cited by 1
A Pose and Style-Invariant Reenactment Technique for Artistic Portraits Using GAN
Oh, J., Min, K., & Yang, H. (2022) · IEEE Access · doi:10.1109/access.2022.3172297 · Cited by 2
Facial Attribute Editing based on Independent Selective Transfer Unit and Self-attention Mechanism
Liu, X., Guo, P., Liu, J., Tuo, D., Lei, S., & Wang, Y. (2022) · International Congress on Image and Signal Processing (CISP) · doi:10.1109/cisp-bmei56279.2022.9979903 · Cited by 1
“StarGAN [5] connects the target label and the source image together at the input layer of the generator and applies residual block [10] for image translation, while AttGAN [4] first extracts the high-level semantic information of the source image by using the encoder, and then connects the high-level semantic informat…”
SemanticGAN: Facial Image Editing with Semantic to Realize Consistency
Luan, X., Yang, N., Fan, H., & Tang, Y. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-18913-5_34 · Cited by 1
Manipulated Face Detection and Localization Based on Semantic Segmentation
Li, G., Zhao, X., Cao, Y., & Hu, C. (2023) · Lecture notes in computer science · doi:10.1007/978-3-031-25115-3_7 · Cited by 1
Eye Contact with the Machine: Gaze Correction in Video Conferencing
Rapoport, R., & Tollmann, V. (2023) · Digitale Gesellschaft · doi:10.1515/9783839462287-011 · Cited by 1
Yang, M.-C., Wong, G.-W., & Chen, M. C. (2023) · ACM Transactions on Knowledge Discovery from Data · doi:10.1145/3634751 · Cited by 1
“A key setting in References [20, 40] is labeling the training data for the discriminator as true only when both output and context are true.”
BackgroundGeneration of High-Quality Realistic Faces with StyleGAN
Ashwini, K., Nagajyothi, D., Ramakrishna, C., & Jyothi, V. (2023) · IEEE Global Conference for Advancement in Technology (GCAT) · doi:10.1109/gcat59970.2023.10353468 · Cited by 1
“In [13], the authors Shengju Qian and Ping Tan, introduce a GAN model called AttGAN that allows users to edit facial attributes in images, such as changing the hairstyle or eye color.”
BackgroundTransferability of CNN models for GAN-generated face detection
Aieprasert, T., Mahdlang, Y., Pansiri, C., Sae-Bae, N., & Khomkham, B. (2024) · Multimedia Tools and Applications · doi:10.1007/s11042-024-18664-4 · Cited by 2
“Assuch,theAttGANmodel[7]offersawiderangeof possibilitiesforgeneratingconvincingfakefaceimages.”
InfluentialLuo, M., Gokhale, T., Varshney, N., Yang, Y., & Baral, C. R. (2024) · Synthesis lectures on computer vision · doi:10.1007/978-3-031-57816-8_4 · Cited by 1
Fingerprinting Image-to-Image Generative Adversarial Networks
Li, G., Xu, G., Qiu, H., Guo, S., Wang, R., Li, J., Zhang, T., & Lu, R. (2024) · IEEE European Symposium on Security and Privacy · doi:10.1109/eurosp60621.2024.00011 · Cited by 6
“Furthermore, through extensive evaluations across three primary I2I tasks (attribute editing, domain translation, super resolution) utilizing advanced GAN models (e.g., AttGAN [19], StarGAN [11], STGAN [40]), our method demonstrates high versatility and comprehensiveness.”
MethodYücer, Ş., Atapour–Abarghouei, A., Moubayed, N. A., & Breckon, T. P. (2024) · International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn60899.2024.10650732 · Cited by 3
“Whilst, many common controllable image-to-image-based [20], [23], [6] and latent space interpolation-based methods [15], [1] offer ways to control facial attributes, they do not inherently offer a factorised latent space for explicit control over image attributes.”
BackgroundFaceDefend: Copyright Protection to Prevent Face Embezzle
Zhai, R., Ni, R., Yu, Y., & Zhao, Y. (2024) · ACM Transactions on Multimedia Computing Communications and Applications · doi:10.1145/3699718 · Cited by 2
“Some models [6, 19, 26] modify attributes such as expression, hair color, gender, and eyeglasses.”
An improved face attributes editing method based on DDIM
He, L., Chen, Q., Pang, Y., Wang, M., Wu, Y., Liu, L., & Qiang, Z. (2024) · Scientific Reports · doi:10.1038/s41598-024-78378-3 · Cited by 1
SoLAD: Sampling Over Latent Adapter for Few Shot Generation
Mondal, A. K., Tiwary, P., Singla, P., & Prathosh, A. P. (2024) · IEEE Signal Processing Letters · doi:10.1109/lsp.2024.3496822 · Cited by 1
Mitigating the Impact of Attribute Editing on Face Recognition
Banerjee, S., Mullangi, S. P., Wagle, S., Hegde, C., & Memon, N. D. (2024) · IEEE International Joint Conference on Biometrics (IJCB) · doi:10.1109/ijcb62174.2024.10744430 · Cited by 1
“In [1], the authors report ArcFace performance after using AttGAN [14] for attribute editing on 100 subjects from LFW (subjects may not be identical to our set).”
BackgroundMethodControllable facial protection against malicious translation-based attribute editing
Xie, Y., Zhou, Y., Wang, T., Hua, Z., Wen, W., Yi, S., & Zhang, Y. (2024) · Knowledge-Based Systems · doi:10.1016/j.knosys.2024.112873 · Cited by 1
StyleGan for Deepfake Generation: A Comparative Study
Sghir, A. A., Amraoui, S. E., Elafi, I., Zrira, N., & Benmiloud, I. (2025) · Advances in Science, Technology & Innovation/Advances in science, technology & innovation · doi:10.1007/978-3-031-74470-9_13 · Cited by 2
Choi, C., Kim, D., Park, S., Lee, H., Kim, H., Lee, S. M., & Kim, N. (2025) · Computers in Biology and Medicine · doi:10.1016/j.compbiomed.2025.110846 · Cited by 1
Modifying memorability of facial images using attention guided generative adversarial network
Sadana, A., Anand, A., Poria, N., Thakur, N., & Seeja, K. R. (2025) · Franklin Open · doi:10.1016/j.fraope.2025.100360 · Cited by 1
FaceMixup: Enhancing Facial Expression Recognition through Mixed Face Regularization
Souza, M., Faria, F. A., Texeira, R., & Segundo, M. P. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-31933-3_46 · Cited by 1
Face Completion Using Generative Adversarial Network
Raut, P., Doshi, M., Diwan, M., & Doshi, K. (2020) · Algorithms for intelligent systems · doi:10.1007/978-981-15-3242-9_50 · Cited by 1
Pose transfer based on generative adversarial networks
Pan, H., & Cao, X. (2020) · International Conference on Image, Video Processing and Artificial Intelligence (IVPAI) · doi:10.1117/12.2578040 · Cited by 1
“Conditional CycleGAN utilizes CGAN with a cyclic consistency framework, Invertible Conditional GANs (IcGAN) [8] , and AttGAN [9] attempted to transfer visual attributes.”
MethodDialog Driven Face Construction using GANs
Vijay, M., Meghana, M., Aklecha, N., & Srinath, R. (2020) · IEEE International Conference on Tools with Artificial Intelligence (ICTAI) · doi:10.1109/ictai50040.2020.00104 · Cited by 1
“The base face is generated with an MSG-Style GAN [3] [4], while the attribute editing is achieved using the AttGAN architecture [5].”
MethodInfluentialConverting Image Labels to Meaningful and Information-rich Embeddings
Karatsiolis, S., & Kamilaris, A. (2021) · International Conference on Pattern Recognition Applications and Methods · doi:10.5220/0010375801070119 · Cited by 1
“He et al., 2019) uses an encoder-decoder architecture but additionally applies conditional decoding of the latent representation based on the desired attributes (i.e., class labels).”
BackgroundMertes, S., Lingenfelser, F., Kiderle, T., Dietz, M., Diab, L., & André, E. (2021) · International Conference on Deep Learning Theory and Applications (DeLTA) · doi:10.5220/0010549401320139 · Cited by 1
“In contrast to the use-cases tackled by our approach, face editing does not aim to generate completely new data, but to modify existing image data (He et al., 2019; Royer et al., 2020; Choi et al., 2018; Liu et al., 2017; Lin et al., 2018).”
Unsupervised multi-domain image translation with domain representation learning
Liu, H., Chen, L., Sui, H., Zhu, Q., Lei, D., & Liu, S. (2021) · Signal Processing Image Communication · doi:10.1016/j.image.2021.116452 · Cited by 1
Attribute Centered Multimodal Response Generation in a Dialogue System
Firdaus, M., Shandilya, A. P., Pujari, S. P., & Ekbal, A. (2021) · International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn52387.2021.9533312 · Cited by 1
“Similarly, [13] proposed an attribute Generative Adversarial Network (GAN) for incorporating different features in facial images.”
BackgroundMUSE: Textual Attributes Guided Portrait Painting Generation
Hu, X., Yu, P., Knight, K. K., Ji, H., Li, B., & Shi, H. (2021) · IEEE International Conference on Multimedia Information Processing and Retrieval (MIPR) · doi:10.1109/mipr51284.2021.00072 · Cited by 1
“and AttGAN [8] by modifying hair , face expression , mood and gender respectively.”
BackgroundMethodResultInfluentialImage Attribute Migration Based on Decoupling and Adaptive Layer Instance Normalization
Cai, X., Li, F., Chen, K., Wei, Y., & Sun, H. (2022) · International Journal of Pattern Recognition and Artificial Intelligence · doi:10.1142/s0218001422540118 · Cited by 1
VR-FAM: Variance-Reduced Encoder with Nonlinear Transformation for Facial Attribute Manipulation
Yuan, Y., Ma, S., & Zhang, J. (2022) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp43922.2022.9746046 · Cited by 1
Controllable Facial Micro-element Synthesis using Segmentation Maps
Kim, Y., & Park, I. K. (2023) · International Conference on Automatic Face and Gesture Recognition · doi:10.1109/fg57933.2023.10042571 · Cited by 1
“There are also several works of converting input images into styles of other facial images [21] and works of editing certain attributes such as hair, beard, and glasses [6], [10], [24].”
BackgroundUnsupervised Discovery and Manipulation of Continuous Disentangled Factors of Variation
Fontanini, T., Donati, L., Bertozzi, M., & Prati, A. (2023) · ACM Transactions on Multimedia Computing Communications and Applications · doi:10.1145/3591358 · Cited by 1
Can GAN-Induced Attribute Manipulations Impact Face Recognition?
Banerjee, S., Aggarwal, A., & Ross, A. A. (2023) · Lecture notes in computer science · doi:10.1007/978-3-031-37731-0_38 · Cited by 1
Sugar-coated poison defense on deepfake face-swapping attacks
Guo, C.-Y., & Yu, F. (2024) · doi:10.1145/3644032.3644459 · Cited by 3
MGGR: MultiModal-Guided Gaze Redirection with Coarse-to-Fine LearningPreprint
Chen, J., Zhang, J., Fan, J., Chen, T., Sangineto, E., & Sebe, N. (2020) · arXiv (Cornell University) · Cited by 1
Actionable Attribution Maps for Scientific Machine LearningPreprint
Liu, S., Kailkhura, B., Zhang, J., Hiszpanski, A. M., Robertson, E. L., Loveland, D., & Han, T. Y.-J. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2006.16533 · Cited by 1
“After obtaining these attributes, we train the attGAN (He et al., 2019) that allows us to selectively edit the salient attributes of a given SEM image..”
BackgroundMethodInfluentialOne-shot domain adaptation for semantic face editing of real world images using StyleALAEPreprint
Reddy, R. K., Shubham, K., Venkatesh, G. S., Gandikota, S., Khoche, S., Jayagopi, D. B., & Srinivasaraghavan, G. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2108.13876 · Cited by 1
“Researchers have also explored conditional GAN-based model [5, 6, 15] for semantic face editing, but training such a model often requires real world images with necessary semantic variations, which is often hard to obtain.”
Background2019 Index IEEE Transactions on Image Processing Vol. 28
(2019) · IEEE Transactions on Image Processing · doi:10.1109/tip.2019.2959522 · Cited by 1
Local Facial Attribute Transfer through InpaintingPreprint
Durall, R., Pfreundt, F., & Keuper, J. (2020) · Opus-HSO (Offenburg University of Applied Sciences) · doi:10.1109/icpr48806.2021.9412878 · Cited by 2
“Our model ATI-GAN: Attribute Transfer Inpainting Generative Adversarial Network Previous methods [1,2] generating new images (global) Our approach generating new patches (local) ATI-GAN”
MethodEnd-to-End FusVAE for Face Image Fusion
Li, X., Chen, B., Wen, M., & Wang, H. (2018) · International Conference on Algorithms, Computing and Artificial Intelligence (ACAI) · doi:10.1145/3302425.3302438
“put forward Attribute Generative Adversarial Networks (AttGAN) [5], which combine VAE and GAN.”
BackgroundApplications of Generative Adversarial Models in Visual Search ReformulationPreprint
Xiao, K., Hu, H., & Wang, Y. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1910.12460
“AttGAN does manage to capture finer level attribute manipulation in the context of facial editing, but conditions the generator training on attribute classifications and makes few architectural changes to ensure unsupervised separation of high level attributes [5].”
BackgroundSurvey Paper on Person Attribute Changes using GANS
Charles, P. (2020) · International Journal of Engineering Research and · doi:10.17577/ijertv8is120327
Hierarchical neural network for facial attribute transfer
Guo, Y. G., Shang, Y., & Zhang, X. (2020) · IOP Conference Series Materials Science and Engineering · doi:10.1088/1757-899x/768/7/072088 · Cited by 1
“Specifically, if we turn the gender attribute of a short-haired male image into a female, short hair will also become long hair in AttGAN.”
BackgroundMethodInfluentialExemplar-based Generative Facial EditingPreprint
Guo, J., Liu, Y., Qian, Z., & Zhou, Z. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2006.00472
“StarGAN [1] and AttGAN [2] employ the attribute labels as conditional information to generate a new face with desired attributes while preserving other details.”
BackgroundThe pursuit of beauty: Converting image labels to meaningful vectorsPreprint
Karatsiolis, S., & Kamilaris, A. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2008.00665
“The representations with the attributes exchanged passed through a label classifier and a real/fake discriminator. The latter used an adversarial loss aiming to produce more realistic images. AttGan [11] also used an encoder-decoder architecture but additionally applied a conditional decoding of the latent representat…”
MethodOverview of Face Attribute Generation
Chen, L., Zhou, G., Liu, Y., & Jin, J. (2020) · Chinese Control and Decision Conference (CCDC) · doi:10.1109/ccdc49329.2020.9164467 · Cited by 1
“Based on published papers on face attribute generation, it can be roughly divided into specific domains[7] and multi-domains[8] to generate two major categories, each of which will have multiple algorithm models.”
BackgroundMethodAdamiak, K., Kapusta, P., & Ślot, K. (2020) · International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn48605.2020.9207097
“Other approaches are based, for example on applying attribute classifi-cation constraints [40] or by introducing semantic label masks into a training process in order to achieve interactive face editing and manipulation [41].”
MethodImage Modification using Text with GANs
Doshi, F., Doshi, P., Gandhi, J., Dwivedi, K., & Mangrulkar, R. S. (2020) · International Journal of Computer Applications Technology and Research · doi:10.7753/ijcatr0911.1001
“Text Adaptive Discriminators [6]: In order to consider the effect of each word on the text, the idea is to have word-level discriminators.”
BackgroundToward U-Net-based GANs for Diverse Facial Image Synthesis from Sketch
Phusomsai, W., & Limpiyakorn, Y. (2020) · International Conference on Internet Computing for Science and Engineering · doi:10.1145/3424311.3424323
“He et al. [10] proposed AttGAN using latent representation of a face input image concatenated with the target label and passing it through the encoder-decoder architecture for generating the output image.”
Method
Metadata
- Authors
- He, Z., Zuo, W., Kan, M., Shan, S., & Chen, X.
- Affiliations
- Chinese Academy of Sciences (CN)
- Harbin Institute of Technology (CN)
- Venue
- IEEE Transactions on Image Processing
- Published
- 2019-11
- Volume
- 28
- Issue
- 11
- Pages
- 5464-5478
- ISSN
- 1057-7149 (ISSN-L)
- 1941-0042
- OpenAlex
- W2963626105
- Cited by
- 833
- References
- 52
- Citations by year
- 2026: 57 · 2025: 110 · 2024: 113 · 2023: 134 · 2022: 115 · 2021: 158 · 2020: 112 · 2019: 32
- FWCI
- 46.59
- Type
- journal article
- Language
- en
- Open access
- closed
- Licence
- Locations
- IEEE Transactions on Image Processing (publishedVersion)
- PubMed (publishedVersion)
- Funders
- National Key R&D Program of China (2017YFA0700800)
- National Natural Science Foundation of China (61671182, 61772496, 61732004)
- Topics
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Generative Adversarial Networks and Image Synthesis
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Face recognition and analysis
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Digital Media Forensic Detection
- Keywords
- facial attribute editing, generative adversarial networks, encoder-decoder architecture, latent representations, face synthesis, CelebA, Labeled Faces in the Wild
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.