Pith. sign in

REVIEW 3 major objections 5 minor 39 references

Mocycle-GAN: Unpaired Video-to-Video Translation

T0 review · 3 major / 5 minor · reviewed 2026-08-14 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Mocycle-GAN claims that unpaired video translation improves when motion, not just appearance, is cycled and transferred across domains.

desk verdict A solid, honest extension of CycleGAN with motion cycle consistency and a learned motion translator; the pseudo-target supervision for the translator is a real but non-fatal soft spot. read the letter →

arxiv 1908.09514 v1 pith:UNZ6ZJMG submitted 2019-08-26 cs.CV cs.MM

Yang Chen , Yingwei Pan , Ting Yao , Xinmei Tian , Tao Mei This is my paper · ORCID
classification cs.CVcs.MM
keywords unpairedvideo-to-videotranslationmotioncycleconsistencyopticalflowgenerativeadversarialnetworkstemporalcoherencevideosegmentationcycle-consistentGAN
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper tries to extend unpaired image translation between domains to video, where both per-frame appearance and the motion between consecutive frames ought to survive translation. It proposes adding two motion-aware losses to the CycleGAN skeleton: a motion cycle consistency loss that asks an optical-flow estimate to be reconstructible after a round-trip translation, and a motion translation loss that transfers flow across domains and uses the transferred flow to warp synthetic frames into agreement with the next synthetic frame. On the Viper benchmark, the resulting Mocycle-GAN reports better segmentation-based scores than CycleGAN, RecycleGAN, and a style-transfer baseline, and qualitative results on flower and ambient-condition videos support the same conclusion. The practical stake is that unpaired video translation can be made temporally coherent without any paired video data.

What carries the argument

The load-bearing object is the motion translator $M_X, M_Y$ coupled with two motion-guided losses. For a pair of consecutive frames, an off-the-shelf optical flow network provides a flow field; the motion cycle consistency loss penalizes the difference between the flow before the round trip and the flow after it, masked to pixels that are traceable. The motion translation loss sends the source flow through the motion translator, warps the synthetic frame with the translated flow, and compares the warped frame to the actual next synthetic frame at traceable pixels. The motion translator itself is trained on the generator's own synthetic frames, using the flow between them as a pseudo target, and the whole model is optimized by alternating between updating generators/discriminators and updating motion translators in an EM-style schedule.

What would settle it

Take a held-out set of consecutive video pairs from Viper whose ground-truth target optical flow is available, run the trained source-to-target motion translator on source flows, and measure the endpoint error between transferred flows and the true target flows. If the transferred flows are no closer to ground truth than the untranslated source flows, then the motion translator has not learned cross-domain motion and any measured translation gain must come from the cycle-consistency regularization alone.

Watch

Extended reading notes

Core claim

The central claim is that explicitly representing motion as optical flow and constraining it twice -- once round-trip through the translation cycle and once as a cross-domain transfer that warps synthetic frames -- yields unpaired video translation that is both frame-realistic and temporally continuous. The paper states this as: integrating motion estimation into an unpaired video translator improves over frame-level cycle-consistency alone and over prior temporal-prediction schemes. The evidence is the segmentation metrics on video-to-labels and labels-to-video translation on Viper, plus human preference on ambient-condition and flower videos.

Load-bearing premise

The motion translator is trained to imitate flows computed between the generator's own synthetic frames, so the whole temporal-coherence mechanism assumes that those synthetic-frame flows are accurate enough to serve as the target training signal; if the generator is temporally inconsistent, the motion translation loss degenerates into a self-consistency regularizer without real cross-domain motion transfer.

Editorial extensions

If this is right

  • Unpaired video translation can preserve temporal continuity by enforcing that motion survives a translation cycle, not just appearance.
  • The motion translator, once trained, provides a cross-domain flow estimate that can be applied at inference or combined with frame generation to smooth outputs.
  • On Viper, video-to-labels and labels-to-video translation scores improve across mean pixel accuracy, average class accuracy, and IoU relative to the compared baselines.
  • The approach extends to settings without paired supervision where target optical flow is unknown, including flower-to-flower and day-to-night changes.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Because the motion translator is trained only on pseudo-target flows obtained from the generator's synthetic frames, its quality is bounded by the generator's temporal coherence early in training; a direct, cross-domain flow test on paired Viper clips could reveal how much of the gain is from true motion transfer rather than self-consistency regularization.
  • The motion translation constraint can be viewed as a differentiable way to enforce temporal smoothness, and it may also act as a regularizer that prevents the generator from treating frames independently, which would explain why it helps even at inference time when only single-frame translation is used.
  • A natural testable extension is to plug the motion cycle loss into other unpaired translation architectures or to replace the optical-flow estimator with a learned warp predictor, then measure whether the improvements persist on high-motion or heavily occluded videos.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 5 minor

Summary. The paper proposes Mocycle-GAN, an unpaired video-to-video translation method that augments CycleGAN with two motion-guided temporal constraints: motion cycle consistency, which enforces that optical flow is reconstructed through the translation cycle, and motion translation, which uses learned translators MX and MY to map source-domain flow into the target domain and then warps synthetic frames to encourage temporal coherence. The model is trained with an alternating procedure that updates generators/discriminators (with adversarial, frame-cycle, motion-cycle, and motion-translation losses) and motion translators (with an auxiliary loss against 'pseudo' target flows computed from the generators' own outputs). Experiments on Viper for video-to-labels and labels-to-video report improvements over CycleGAN, RecycleGAN, and a CycleGAN variant with direct warping (Cycle-GAN_SF) in mean pixel accuracy, average class accuracy, and IoU; additional qualitative and human-preference experiments cover ambient condition transfer and flower-to-flower translation.

Significance. If the reported gains are robust, Mocycle-GAN presents a sensible way to incorporate explicit motion information into unpaired video translation, going beyond RecycleGAN's implicit temporal prediction. The architecture is clearly described, the ablations support the contribution of each added loss term, and the use of off-the-shelf optical flow (FlowNet) makes the approach practical. However, the central claim that motion translation improves temporal coherence is not directly measured: the reported metrics are per-frame appearance scores, and the motion translators are trained against a self-referential pseudo-flow target. The significance is therefore conditional on additional temporal-consistency evidence and on demonstrating that the pseudo-flow supervision is not circular.

major comments (3)
  1. [Section 3.3, Eq. (4)] The motion translators MX and MY are trained to match the optical flow fext and feys computed between the generators' own synthetic frames. This is a self-referential training signal: if the generators already produce temporally coherent frames, the pseudo-flow is informative, but if they are temporally inconsistent, the flow estimate is corrupted by appearance artifacts, and the auxiliary loss only enforces a circular consistency that does not ground motion in real target dynamics. The claim that motion translation strengthens temporal continuity therefore needs an independent check. Please provide a quantitative temporal-consistency metric on the synthetic videos (e.g., warping error of consecutive synthetic frames against real target flow, or a flicker measure) and, ideally, an analysis of how the pseudo-flow target evolves during training. Absent this, the motion-translation constraint may amount to a learned self-warping that does not improve temporal coherence.
  2. [Table 3 and Table 4, ablation study] The incremental gain of the motion-translation component (Cycle-GAN + MT) over the direct-warping baseline Cycle-GAN_SF is small: e.g., 63.0 vs. 62.3 in MP for video-to-labels and 47.0 vs. 44.6 in MP for labels-to-video, with similar small gaps in AC and IoU. No standard deviations, confidence intervals, or significance tests are reported for any of the ablation numbers, so it is unclear whether this difference is meaningful or within run-to-run variation. Since the paper's central contribution is the motion translator beyond simple warping, please report multiple runs with error bars or a paired significance test (e.g., per-video scores).
  3. [Section 4, evaluation metrics] All quantitative metrics (MP, AC, IoU, and FCN scores) are per-frame appearance or segmentation measures; none directly evaluates temporal consistency, which is the stated motivation for the motion cycle consistency and motion translation constraints. The paper needs to include at least one temporal coherence metric (e.g., flow warping error between consecutive synthetic frames, temporal frame interpolation error, or a user study that specifically asks about temporal artifacts such as flicker) to substantiate the claim that Mocycle-GAN produces temporally continuous videos. Without such a metric, the reported gains could in principle reflect improved per-frame appearance rather than improved motion modeling.
minor comments (5)
  1. [Abstract and throughout] The word 'constrain' is used where 'constraint' is intended (e.g., 'three types of constrains' in the abstract and 'spatial/temporal constrains' in Section 3). Please correct these throughout the manuscript.
  2. [Eq. (5) and surrounding text] The warping operation is described as 'bi-linear interpolation'; the standard spelling is 'bilinear'. Also, the notation W(efxt, ext) and W(efys, eys) appears with inconsistent spacing and subscript formatting; please make the warping operator definition and its arguments unambiguous.
  3. [Algorithm 1, line 6] The optical-flow variables are typeset with unusual spacing (e.g., 'fx r e ct'), which appears to be a LaTeX formatting issue. Please correct the subscripts for fx_rec and fyr_rec so they are readable.
  4. [Section 4.1, Implementation Details] The loss weights are written as λF C, λMC, and λMT with spaces; these should be λ_FC, λ_MC, and λ_MT. The paper also does not state the number of training iterations N or the convergence criterion used in Algorithm 1; adding this would improve reproducibility.
  5. [Section 4.1, Compared Approaches] The list of compared methods has a numbering error: two items are labeled '(4)'. The second '(4)' should be '(5)' for Mocycle-GAN. Please fix the enumeration.

Circularity Check

1 steps flagged · score 6.0 of 10

Motion translator is trained on pseudo target flows from the generator's own synthetic frames, so the transferred-flow prediction reduces to self-warping consistency rather than independent motion transfer.

  1. fitted input called prediction [Section 3.3, Eq. (4) and Eq. (5)]
    "Here we directly utilize the optical flow fext between the corresponding synthetic frames in target domain as the “pseudo” target optical flow for training motion translator. ... the auxiliary motion consistency loss is defined as L1 distance between the transferred optical flow and “pseudo” target optical flow: LAM(MX , MY) = sum_t ||efxt - fext||_1 + sum_s ||efys - feys||_1."

    The supervision for motion translator MX is the optical flow fext computed by FlowNet between the generator's own synthetic frames ext and ext+1. Eq. (4) fits MX to output exactly that pseudo flow. Eq. (5) then penalizes the mismatch between W(MX(fxt), ext) and ext+1. When Eq. (4) is minimized, MX(fxt) approximates F(ext, ext+1), which by definition is the flow that warps ext to ext+1. Thus Eq. (5) reduces to a self-consistency constraint on the generator's own outputs: it forces consecutive synthetic frames to be mutually warp-consistent, but it never uses a real target-domain flow. The claimed 'transfer of motion across domains' is therefore a fitted self-referential quantity, not an independent prediction validated against target motion.

full rationale

The frame-level part of Mocycle-GAN is self-contained and independently grounded: the adversarial loss and frame cycle consistency are standard Cycle-GAN components, and the reported MP/AC/IoU scores are evaluated against external segmentation ground truth. No load-bearing self-citation chain was found; the cited prior work is used for architecture or comparison, not to establish the central claim. However, the paper's main novelty, the motion translation constraint, is trained with 'pseudo' target optical flows computed from the generator's own synthetic frames (Eq. 4). Because the same synthetic frames are used in the motion translation warping loss (Eq. 5), the motion translator is effectively fitted to reproduce the generator's self-flow, making the 'transferred motion' a circular self-consistency signal rather than an independent transfer of source motion to real target dynamics. The qualitative claim in Figure 6 that transferred flow matches ground-truth target flow is not supported by training on real target flows. This is a partial circularity: the overall system still has independent frame-level content, but the central motion-transfer component reduces to a fitted self-consistency constraint.

Assumptions & free parameters 1 free parameters · 4 assumptions · 1 invented entities

The core new component, the motion translator, rests on an unverified surrogate target: the flow of generated frames is used as a stand-in for real target motion. The remaining assumptions are standard within the CycleGAN literature but are not independently verified here.

free parameters (1)
  • Loss weights lambda_FC, lambda_MC, lambda_MT = 10, 10, 10
    Set by hand without sensitivity analysis; these weights balance frame cycle consistency, motion cycle consistency, and motion translation in Eq. (6).
assumptions (4)
  • domain assumption FlowNet provides accurate optical flow for real, synthetic, and reconstructed frames
    All motion constraints depend on FlowNet outputs, shown in Section 3.1 and used in Eq. (3) and Eq. (5).
  • standard math Visibility masks C_xt from [11] correctly identify occluded and boundary pixels
    Used in Eq. (3) and Eq. (5); adopted without re-validation in the target domains.
  • ad hoc to paper The optical flow between the generator's synthetic frames is a good surrogate for real target motion
    Eq. (4) defines the pseudo-target using f_ext, the flow of the model's own outputs; no external target flow is used during training.
  • domain assumption CycleGAN's cycle consistency aligns appearance across domains
    Used as the backbone for frame-level translation, inherited from [37].
invented entities (1)
  • Motion translators MX and MY
    purpose: Map source-domain optical flow to target-domain optical flow and warp synthetic frames to enforce temporal consistency.
    New network modules trained solely with pseudo-targets from the generator's own synthetic frames; no external validation data is used.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Mocycle-GAN: Unpaired Video-to-Video Translation." pith.science (2026). https://pith.science/paper/UNZ6ZJMG

@misc{pith2026190809514,
  author       = {Pith},
  title        = {Pith review of: Mocycle-GAN: Unpaired Video-to-Video Translation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/UNZ6ZJMG}},
  note         = {Machine review of arXiv:1908.09514}
}
read the original abstract

Unsupervised image-to-image translation is the task of translating an image from one domain to another in the absence of any paired training examples and tends to be more applicable to practical applications. Nevertheless, the extension of such synthesis from image-to-image to video-to-video is not trivial especially when capturing spatio-temporal structures in videos. The difficulty originates from the aspect that not only the visual appearance in each frame but also motion between consecutive frames should be realistic and consistent across transformation. This motivates us to explore both appearance structure and temporal continuity in video synthesis. In this paper, we present a new Motion-guided Cycle GAN, dubbed as Mocycle-GAN, that novelly integrates motion estimation into unpaired video translator. Technically, Mocycle-GAN capitalizes on three types of constrains: adversarial constraint discriminating between synthetic and real frame, cycle consistency encouraging an inverse translation on both frame and motion, and motion translation validating the transfer of motion between consecutive frames. Extensive experiments are conducted on video-to-labels and labels-to-video translation, and superior results are reported when comparing to state-of-the-art methods. More remarkably, we qualitatively demonstrate our Mocycle-GAN for both flower-to-flower and ambient condition transfer.

Figures

Figures reproduced from arXiv: 1908.09514 by the authors.

Figure 1
Figure 1. Given any two unpaired video collections X and Y , our Mocycle-GAN learns to translate videos from source domain X to target domain Y . Left: Translation between Game scene videos and segmentation label maps. Center: Translation between time-lapse videos of variant flowers. Right: Translation between Game scene videos under different ambient conditions, e.g., rendering day-light video to the sunset/night environment… view at source ↗
Figure 2
Figure 2. Comparison between two unpaired translation ap￾proaches and our Mocycle-GAN. (a) Cycle-GAN exploits cycle￾consistency constraint to model appearance structure for unpaired image-to-image translation. (b) Recycle-GAN utilizes temporal pre￾dictor (PX and PY ) to explore cycle consistency across both domains and time for unpaired video-to-video translation. (c) Mocycle-GAN explicitly models motion across frames with op… view at source ↗
Figure 3
Figure 3. The overview of Mocycle-GAN for unpaired video-to-video translation (X: source domain; Y : target domain). Note that here we only depict the forward cycle X → Y → X for simplicity. Mocycle-GAN consists of generators (GX and GY ) to synthesize frames across domains, discriminators (DX and DY ) to distinguish real frames from synthetic ones, and motion translator (MX ) for motion translation across domains. Given two … view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Examples of (a) video-to-labels and (b) labels-to-video results in Viper dataset under various ambient conditions. The original inputs, the output results by different models, and the ground truth outputs are given [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]
Figure 5
Figure 5. Figure 5: Examples of night-to-day results in Viper dataset. The original inputs and the output results by different models are given. Each row denotes one sequence of frames. Source Frame Overlay Source Optical Flow Transferred Optical Flow Target Optical Flow Target Frame Over…
Figure 6
Figure 6. Figure 6: Examples of motion translation results in video-to-labels. From left to right: Source frame overlay, optical flow in source, transferred optical flow via motion translator, ground truth optical flow in target, and ground truth target frame overlay. Moreover, to fully v…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

39 extracted references · 37 canonical work pages

  1. [1]

    Alexander G Anderson, Cory P Berg, Daniel P Mossing, and Bruno A Olshausen

    work page

  2. [2]

    Aayush Bansal, Shugao Ma, Deva Ramanan, and Yaser Sheikh. 2018. Recycle-gan: Unsupervised video retargeting. In ECCV

    work page 2018

  3. [3]

    Dongdong Chen, Jing Liao, Lu Yuan, Nenghai Yu, and Gang Hua. 2017. Coherent online video style transfer. In ICCV

    work page 2017

  4. [4]

    Yunjey Choi, Minje Choi, Munyoung Kim, Jung-Woo Ha, Sunghun Kim, and Jaegul Choo. 2018. Stargan: Unified generative adversarial networks for multi- domain image-to-image translation. In CVPR

    work page 2018

  5. [5]

    Chang Gao, Derun Gu, Fangjun Zhang, and Yizhou Yu. 2018. ReCoNet: Real-time Coherent Video Style Transfer Network. arXiv preprint arXiv:1807.01197 (2018)

    work page Pith review arXiv 2018

  6. [6]

    Leon A Gatys, Alexander S Ecker, and Matthias Bethge. 2016. Image style transfer using convolutional neural networks. In CVPR

    work page 2016

  7. [7]

    Golnaz Ghiasi, Honglak Lee, Manjunath Kudlur, Vincent Dumoulin, and Jonathon Shlens. 2017. Exploring the structure of a real-time, arbitrary neural artistic stylization network. In BMVC

    work page 2017

  8. [8]

    Ian Goodfellow, Jean Pouget-Abadie, Mehdi Mirza, Bing Xu, David Warde-Farley, Sherjil Ozair, Aaron Courville, and Yoshua Bengio. 2014. Generative adversarial nets. In NIPS

    work page 2014

Show all 39 references
  1. [9]

    Agrim Gupta, Justin Johnson, Alexandre Alahi, and Li Fei-Fei. 2017. Characteriz- ing and improving stability in neural style transfer. In ICCV

    2017

  2. [10]

    Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. 2016. Deep residual learning for image recognition. In CVPR

    2016

  3. [11]

    Haozhi Huang, Hao Wang, Wenhan Luo, Lin Ma, Wenhao Jiang, Xiaolong Zhu, Zhifeng Li, and Wei Liu. 2017. Real-time neural style transfer for videos. InCVPR

    2017

  4. [12]

    Eddy Ilg, Nikolaus Mayer, Tonmoy Saikia, Margret Keuper, Alexey Dosovitskiy, and Thomas Brox. 2017. Flownet 2.0: Evolution of optical flow estimation with deep networks. In CVPR

    2017

  5. [13]

    Phillip Isola, Jun-Yan Zhu, Tinghui Zhou, and Alexei A Efros. 2017. Image-to- image translation with conditional adversarial networks. In CVPR

    2017

  6. [14]

    Justin Johnson, Alexandre Alahi, and Li Fei-Fei. 2016. Perceptual losses for real-time style transfer and super-resolution. In ECCV

    2016

  7. [15]

    Taeksoo Kim, Moonsu Cha, Hyunsoo Kim, Jung Kwon Lee, and Jiwon Kim. 2017. Learning to discover cross-domain relations with generative adversarial networks. In ICML

    2017

  8. [16]

    Diederik P Kingma and Jimmy Ba. 2014. Adam: A method for stochastic opti- mization. In ICLR

    2014

  9. [17]

    Yehao Li, Ting Yao, Yingwei Pan, Hongyang Chao, and Tao Mei. 2018. Jointly localizing and describing events for dense video captioning. In CVPR

    2018

  10. [18]

    Ming-Yu Liu and Oncel Tuzel. 2016. Coupled generative adversarial networks. In NIPS

    2016

  11. [19]

    Jonathan Long, Evan Shelhamer, and Trevor Darrell. 2015. Fully convolutional networks for semantic segmentation. In CVPR

    2015

  12. [20]

    Youssef Alami Mejjati, Christian Richardt, James Tompkin, Darren Cosker, and Kwang In Kim. 2018. Unsupervised Attention-guided Image-to-Image Translation. In NIPS

    2018

  13. [21]

    Yingwei Pan, Yehao Li, Ting Yao, Tao Mei, Houqiang Li, and Yong Rui. 2016. Learning Deep Intrinsic Video Representation by Exploring Temporal Coherence and Graph Structure. In IJCAI

    2016

  14. [22]

    Yingwei Pan, Tao Mei, Ting Yao, Houqiang Li, and Yong Rui. 2016. Jointly modeling embedding and translation to bridge video and language. In CVPR

    2016

  15. [23]

    Yingwei Pan, Zhaofan Qiu, Ting Yao, Houqiang Li, and Tao Mei. 2017. To create what you tell: Generating videos from captions. In ACM MM

    2017

  16. [24]

    Yingwei Pan, Ting Yao, Houqiang Li, and Tao Mei. 2017. Video captioning with transferred semantic attributes. In CVPR

    2017

  17. [25]

    Adam Paszke, Sam Gross, Soumith Chintala, Gregory Chanan, Edward Yang, et al. 2017. Automatic differentiation in PyTorch. In NIPS Workshop

    2017

  18. [26]

    Zhaofan Qiu, Yingwei Pan, Ting Yao, and Tao Mei. 2017. Deep semantic hashing with generative adversarial networks. In SIGIR

    2017

  19. [27]

    Stephan R Richter, Zeeshan Hayder, and Vladlen Koltun. 2017. Playing for benchmarks. In ICCV

    2017

  20. [28]

    Manuel Ruder, Alexey Dosovitskiy, and Thomas Brox. 2016. Artistic style transfer for videos. In GCPR

    2016

  21. [29]

    Dmitry Ulyanov, Vadim Lebedev, Andrea Vedaldi, and Victor S Lempitsky. 2016. Texture Networks: Feed-forward Synthesis of Textures and Stylized Images.. In ICML

    2016

  22. [30]

    Carl Vondrick, Hamed Pirsiavash, and Antonio Torralba. 2016. Generating videos with scene dynamics. In NIPS

    2016

  23. [31]

    Ting-Chun Wang, Ming-Yu Liu, Jun-Yan Zhu, Guilin Liu, Andrew Tao, Jan Kautz, and Bryan Catanzaro. 2018. Video-to-video synthesis. In NIPS

    2018

  24. [32]

    Ting-Chun Wang, Ming-Yu Liu, Jun-Yan Zhu, Andrew Tao, Jan Kautz, and Bryan Catanzaro. 2018. High-resolution image synthesis and semantic manipulation with conditional gans. In CVPR

    2018

  25. [33]

    Xingxing Wei, Jun Zhu, Sitong Feng, and Hang Su. 2018. Video-to-video transla- tion with global temporal consistency. In ACM MM

    2018

  26. [34]

    Xuewen Yang, Dongliang Xie, and Xin Wang. 2018. Crossing-Domain Gener- ative Adversarial Networks for Unsupervised Multi-Domain Image-to-Image Translation. In ACM MM

    2018

  27. [35]

    Zili Yi, Hao Zhang, Ping Tan, and Minglun Gong. 2017. Dualgan: Unsupervised dual learning for image-to-image translation. In ICCV

    2017

  28. [36]

    Hang Zhang and Kristin Dana. 2018. Multi-style generative network for real-time transfer. In ECCV

    2018

  29. [37]

    Jun-Yan Zhu, Taesung Park, Phillip Isola, and Alexei A Efros. 2017. Unpaired image-to-image translation using cycle-consistent adversarial networks. InICCV

    2017

  30. [38]

    Jun-Yan Zhu, Richard Zhang, Deepak Pathak, Trevor Darrell, Alexei A Efros, Oliver Wang, and Eli Shechtman. 2017. Toward multimodal image-to-image translation. In NIPS

    2017

  31. [2016]

    arXiv preprint arXiv:1605.08153 (2016)

    Deepmovie: Using optical flow and deep neural networks to stylize movies. arXiv preprint arXiv:1605.08153 (2016)

    2016 arXiv

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.