Pith. sign in

REVIEW 3 major objections 4 minor 48 references

Time-variant Image Inpainting via Interactive Distribution Transition Estimation

T0 review · 3 major / 4 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read The paper introduces TAMP, a reference-guided inpainting setting where the reference image is time-shifted and possibly damaged, and proposes InDiTE-Diff, which treats complementation as a distribution transition and reports consistent…

desk verdict New task and dataset worth a look; the InDiTE-Diff 'consistent superiority' claim does not survive its own tvRefInpaint table. read the letter →

arxiv 2506.23461 v1 pith:PFVUZ4YO submitted 2025-06-30 cs.CV cs.AI

Yun Xing , Qing Guo , Xiaoguang Li , Yihao Huang , Xiaofeng Cao , Di Lin , Ivor Tsang , Lei Ma This is my paper
classification cs.CVcs.AI
keywords time-variantimageinpaintingreference-guideddiffusionmodeldistributiontransitionsemanticpredictivefilteringconfidencemaskTAMP-Streetdataset
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

The paper introduces Time-vAriant iMage inPainting (TAMP), a task in which a damaged target image must be restored using a reference image of the same scene captured at a different time, so appearance and even objects differ, and the reference may itself be damaged. It argues that existing reference-guided inpainting methods fail on this task because they copy reference content naively, importing semantic contradictions. The paper's solution is InDiTE, a module that treats the two images as samples from geometrically related distributions and interactively complements them, filtering for semantic consistency and producing a confidence mask that tells a diffusion model where to regenerate rather than borrow. Building on a diffusion null-space model, InDiTE-Diff adds a low-pass cross-reference step during sampling so both restored images agree. If the claims hold, damaged pairs of photos of the same scene can be restored even when no clean, contemporaneous reference exists.

What carries the argument

The load-bearing object is the Interactive Distribution Transition Estimation (InDiTE) module, a parameter-shared Siamese U-Net with two heads. It treats the two time-variant images as draws from distributions that share geometry, and learns the transition between them by merging latent features, applying semantic predictive filtering (SPF) to keep only semantically consistent neighbors, and outputting both a complemented image and a confidence mask. The confidence mask is what converts the problem into a standard masked inpainting task: low-confidence regions are handed to a diffusion model, and the low-pass cross-reference term in the sampling loop enforces mutual consistency. The named identity doing the work is the distribution-transition framing: instead of aligning pixels, the module estimates what content can be trusted across temporal change.

What would settle it

Evaluate InDiTE-Diff on time-variant pairs that include strong viewpoint shifts or structural scene changes, such as a building added or removed, where the 'near-identical geometry' premise breaks; if its PSNR advantage over LeftRefill disappears or reverses on those pairs, the central assumption is refuted.

Watch

Extended reading notes

Core claim

The central claim is that time-variant image pairs share near-identical overall geometry, so their complementation can be learned as a distribution transition rather than as explicit copy-and-paste. InDiTE learns this transition with a Siamese U-Net backbone that merges features and applies semantic predictive filtering to keep only consistent contents, while a confidence head learns where the complementation is unreliable. InDiTE-Diff then feeds the complemented, confidence-masked images into a diffusion sampler whose null-space constraint preserves known pixels and whose low-pass cross-reference term forces the two outputs to agree on low frequencies at every step. The paper reports that this consistently outperforms state-of-the-art reference-guided methods on the new TAMP-Street benchmark when both images are damaged, with PSNR gains over LeftRefill of roughly 3.1 to 5.2 dB depending on mask ratio, and remains competitive when the reference is intact.

Load-bearing premise

The whole pipeline assumes the two images of the same scene remain near-identical in overall geometry despite the time gap, so that feature merging, semantic filtering, and low-pass alignment can find genuinely corresponding content.

Editorial extensions

If this is right

  • In tvDuoInpaint, where the reference is also damaged, InDiTE-Diff reports PSNR gains of roughly 3.1 to 5.2 dB over LeftRefill across mask ratios 20% to 60%.
  • InDiTE's complement output can be plugged into existing inpainting models; the paper shows it generally boosts TransFill, TransRef, and LeftRefill for duo-image inpainting.
  • In tvRefInpaint, InDiTE-Diff is competitive with LeftRefill and improves most at higher mask ratios, suggesting the approach extracts more reference value when damage is severe.
  • Ablation shows both InDiTE and the diffusion cross-reference step contribute, with InDiTE providing the majority of the gain.

Reading between the lines

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

  • The confidence mask learned by InDiTE could serve as a general reliability signal for fusing multiple temporally shifted views, beyond the specific inpainting head used here.
  • Because the method weakens the requirement that a reference be clean and contemporaneous, it opens a direct path to restoring scans of archived or historical imagery, though the paper does not test that scenario.
  • The geometric-alignment assumption suggests a natural stress test: applying the method to pairs with camera viewpoint changes would reveal how far the distribution-transition framing can be pushed.
  • If the low-pass cross-reference term is doing most of the consistency work, the method could generalize to video restoration by aligning successive frames in the same latent space; the paper does not explore this.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 4 minor

Summary. The paper introduces Time-vAriant iMage inPainting (TAMP), a reference-guided inpainting setting in which the reference image is captured with a large time gap and may itself be damaged. The authors propose InDiTE, a Siamese U-Net with semantic predictive filtering and a confidence head that produces complemented images and confidence masks, and InDiTE-Diff, which combines InDiTE with the DDNM diffusion model and adds a low-pass cross-reference constraint during sampling. They assemble a TAMP-Street benchmark from VL-CMU-CD images and irregular masks, and report experiments for two settings: tvRefInpaint (intact reference) and tvDuoInpaint (both images damaged). The tvDuoInpaint results show large PSNR gains over LeftRefill, while the tvRefInpaint results are mixed, with LeftRefill winning several low-mask-ratio cells.

Significance. If the reported tvDuoInpaint gains are statistically robust, the paper makes a useful contribution by defining a practical task, providing a plug-and-play complementation module that improves several baselines, and assembling a publicly released benchmark from existing sources. The ablation study in Table V gives some evidence that both the InDiTE complementation and the diffusion cross-reference contribute to the final performance. However, the central advertised claim of consistent superiority over state-of-the-art methods is not supported by the tvRefInpaint results, and the absence of error bars or significance tests leaves the magnitude of even the large tvDuoInpaint gains unquantified. The method also rests on a near-identical-geometry assumption that is not validated on the benchmark.

major comments (3)
  1. [Abstract and Section VI-B, Table II] The abstract's claim that InDiTE-Diff 'consistently outperform SOTA reference-guided image inpainting methods for solving TAMP' is contradicted by the only quantitative tvRefInpaint evidence. In Table II, LeftRefill outperforms InDiTE-Diff on all four metrics at mask ratios 20%-30% (PSNR 30.9928 vs 30.9782; SSIM 0.9438 vs 0.9338) and 30%-40% (PSNR 30.3511 vs 29.9718; SSIM 0.9229 vs 0.9071). The text in Section VI-B, first item, that InDiTE-Diff 'demonstrates a consistent advantage over the baselines' is therefore inaccurate. Please narrow the claim to tvDuoInpaint, or provide paired statistical evidence that the high-mask-ratio tvRefInpaint gains are real and outweigh the low-mask-ratio losses.
  2. [Tables II-V] No standard deviations, confidence intervals, or paired significance tests are reported for any quantitative result. Several comparisons in Table II differ by less than 0.05 dB (e.g., 20%-30%: 30.9928 vs 30.9782), and even the larger tvDuoInpaint gains in Table III cannot be distinguished from run-to-run noise without variance estimates. Please report error bars and, where relevant, paired tests, or explicitly phrase the conclusions as observed point estimates rather than consistent improvements.
  3. [Section IV-A and Eq. (8)] The method is justified by the assumption that time-variant images share 'near-identical overall geometry,' and the semantic predictive filtering in Eq. (3) and the low-pass cross-reference alignment in Eq. (8) both depend on this correspondence. However, TAMP-Street is described in Table I as including Geometric Misalignment as a challenge, and the paper provides no quantitative characterization of alignment in the dataset and no experiment testing robustness to misalignment. This is a correctness-risk concern for the general TAMP claim: if geometric misalignment is substantial, the feature matching and low-pass alignment could enforce wrong correspondences and mislead the diffusion output. Please either restrict the claim to approximately aligned time-variant images or add an experiment that varies or measures misalignment.
minor comments (4)
  1. [Section III-B and Fig. 4 caption] There are typos: 'tuDuoInpaint' should be 'tvDuoInpaint', and the method name is inconsistently written as 'InTiDE-Diff', 'InDiTE-DDNM', and 'ITDiff' in Section VI-E; please standardize.
  2. [Section VI-B and VI-C] PSNR differences are reported as percentages (e.g., '1.3693%', '3.095%/3.4442%'), but PSNR is measured in decibels; the reported quantities are dB differences, not percentage improvements. Please correct the units throughout.
  3. [Section VI-A] TransFill is evaluated by sending data to the authors' server, whereas TransRef and LeftRefill are retrained on TAMP-Street; this asymmetry should be explicitly stated as a potential source of unfair comparison, especially in Table IV where TransFill is used in boosting experiments.
  4. [Section IV-C and Eq. (10)] Several free parameters are not specified: the binarization threshold tau in Eq. (6), the low-pass scale factor D and step size omega in Eq. (8), and the diffusion sampling steps. In addition, the confidence mask is trained against the ground-truth residual C*_i = I*_i - \tilde I_i, so Section VI-E's description of it as an independent indicator should be softened to reflect that this is a supervised regressor fitted to training data.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: InDiTE-Diff's components are supervised or externally sourced, and the central claim is tested against independent baselines; the 'consistently outperform' wording is a correctness concern, not a circular reduction.

full rationale

Walking the claimed derivation chain: TAMP is a task definition (Eq. 1); an empirical study motivates learned complementation; InDiTE learns complementation outputs and confidence masks through supervised losses against ground truth (Eqs. 9-10); InDiTE-Diff wraps DDNM with a low-pass cross-reference term (Eqs. 7-8); and the full system is evaluated on the assembled TAMP-Street benchmark against external baselines TransFill, TransRef, and LeftRefill. No step equates an output to an input by construction. The complementation output ~I_i and confidence map ~C_i are trained with L1/GAN/style/perceptual losses; the confidence target ~C*_i = I*_i - ~I_i is a supervised label, not a fitted parameter renamed as a prediction, and the final PSNR/SSIM/LPIPS values are computed against held-out ground truth. The SPF technique and loss hyperparameters are inherited from MISF [6], a same-group prior paper, but MISF is a peer-reviewed, code-released CVPR method used as a component rather than as a self-citation chain that forces the TAMP result; no uniqueness theorem is imported from the authors' prior work, and DDNM [13] and the low-pass interaction [41] are external. The abstract's 'consistently outperform' wording is in tension with Table II, where LeftRefill wins some tvRefInpaint cells at 20%-40% mask ratios, and no significance tests are reported; that is an evidence/claim mismatch belonging to correctness risk, not circularity. The geometric-similarity assumption in Sec. IV-A is an explicit modeling assumption, not a derived or self-cited fact. I therefore find no significant circularity.

Assumptions & free parameters 3 free parameters · 4 assumptions · 1 invented entities

The central claim rests on an assumed geometric correspondence between time-variant images and on the transferability of SPF-style filtering; three key hyperparameters (tau, D, loss weights) are taken from heuristics or prior work, with tau and D unreported.

free parameters (3)
  • Confidence binarization threshold tau
    Eq. (6) binarizes the confidence map with threshold tau; the value is never reported, which directly affects which pixels are regenerated by the diffusion model.
  • Low-pass filter scale factor D
    Eq. (8) uses low-pass filter phi_D with scale factor D for cross-reference interaction; the value is not given in the paper.
  • Loss weights lambda1..lambda4 = 1, 0.1, 0.1, 250
    Set following MISF [6]; not tuned on TAMP-Street, so the central method depends on these fixed weights.
assumptions (4)
  • domain assumption Time-variant images share near-identical overall geometry
    Section IV-A states this as the basis for modeling complementation as a distribution transition; if violated by viewpoint change or scene rearrangement, SPF feature matching fails.
  • domain assumption Semantic predictive filtering learned from MISF transfers to time-variant pairs
    InDiTE relies on SPF kernels (Eqs. 3-5) to filter inconsistent contents; this is assumed to work across temporal appearance changes.
  • standard math DDNM null-space diffusion is a valid generative prior for the masked complementation inputs
    The method builds on DDNM [13] without modification; correctness of the sampling procedure is assumed from that prior work.
  • domain assumption TAMP-Street masks and image pairs represent the real TAMP distribution
    The benchmark is assembled from VL-CMU-CD street images and randomly paired irregular masks; no evidence this covers other scenes or damage types.
invented entities (1)
  • InDiTE (Interactive Distribution Transition Estimation module)
    purpose: To learn semantic-consistent complementation between time-variant images and predict a confidence mask indicating where complementation is unreliable.
    A learned neural module; no independent theoretical derivation or out-of-benchmark validation; effectiveness is shown only on TAMP-Street.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Time-variant Image Inpainting via Interactive Distribution Transition Estimation." pith.science (2026). https://pith.science/paper/PFVUZ4YO

@misc{pith2026250623461,
  author       = {Pith},
  title        = {Pith review of: Time-variant Image Inpainting via Interactive Distribution Transition Estimation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/PFVUZ4YO}},
  note         = {Machine review of arXiv:2506.23461}
}
read the original abstract

In this work, we focus on a novel and practical task, i.e., Time-vAriant iMage inPainting (TAMP). The aim of TAMP is to restore a damaged target image by leveraging the complementary information from a reference image, where both images captured the same scene but with a significant time gap in between, i.e., time-variant images. Different from conventional reference-guided image inpainting, the reference image under TAMP setup presents significant content distinction to the target image and potentially also suffers from damages. Such an application frequently happens in our daily lives to restore a damaged image by referring to another reference image, where there is no guarantee of the reference image's source and quality. In particular, our study finds that even state-of-the-art (SOTA) reference-guided image inpainting methods fail to achieve plausible results due to the chaotic image complementation. To address such an ill-posed problem, we propose a novel Interactive Distribution Transition Estimation (InDiTE) module which interactively complements the time-variant images with adaptive semantics thus facilitate the restoration of damaged regions. To further boost the performance, we propose our TAMP solution, namely Interactive Distribution Transition Estimation-driven Diffusion (InDiTE-Diff), which integrates InDiTE with SOTA diffusion model and conducts latent cross-reference during sampling. Moreover, considering the lack of benchmarks for TAMP task, we newly assembled a dataset, i.e., TAMP-Street, based on existing image and mask datasets. We conduct experiments on the TAMP-Street datasets under two different time-variant image inpainting settings, which show our method consistently outperform SOTA reference-guided image inpainting methods for solving TAMP.

Figures

Figures reproduced from arXiv: 2506.23461 by the authors.

Figure 1
Figure 1. Illustration of the reference and target image pairs (I1, I2) for (upper) existing RefInpaint studies and (lower) our time-variant image inpainting setting. Gray regions represent the image damages that need to be inpainted. image inpainting methods [5] [6] [7] still struggle to recover the target image when the damage is extensive, especially for those regions with complex semantic structures. To tackle such “ill-p… view at source ↗
Figure 2
Figure 2. Visualization of image inpainting results with existing RefInpaint methods and our method for the tvDuoInpaint case in [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 3
Figure 3. Three main challenges of the proposed time-variant image inpainting tasks. The regions enclosed by purple lines refer to the pixel missing damages. two folds as exemplified in [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗
Figures from the paper (11 more)
Figure 4
Figure 4. Figure 4: LeftRefill results of exemplar time-variant images under tvRefInpaint and tuDuoInpaint setups. Green rectangles indicate the regional ground-truth. B. Empirical Analysis Despite the challenges highlighted above, existing reference￾guided image inpainting methods can st…
Figure 5
Figure 5. Figure 5: Visualization of naive complementation for the time-variant images. The workflow of the two images are highlighted in yellow and blue dashed line respectively. Note that the masks are inversed for a better view [PITH_FULL_IMAGE:figures/full_fig_p004_5.png]
Figure 6
Figure 6. Figure 6: LeftRefill inpainting results with naively complemented time-variant image pair under the tvDuoInpaint setup. successfully preserved in the final “Results”. Nevertheless, it also can be observed that the significant border artifacts caused by naive complementation are …
Figure 7
Figure 7. Figure 7: Technical pipelne of the proposed interactive distribution transition estimation module. To make the process be clear, the data flow of the two images are highlighted with orange and blue arrow lines respectively. The color of mask Mi and confidence mask Ci are inverse…
Figure 8
Figure 8. Figure 8: Illustration of representative time-variant image pairs and mask pairs at varying levels of degradation, adopted to construct our TAMP-Street dataset. further generation. Similar to the complement head, we keep using semantic predictive filter for the confidence head a…
Figure 9
Figure 9. Figure 9: Visualization of existing reference-guided image inpainting methods and our InDiTE-Diff for inpainting typical target time-variant images under tvDuoInpaint setup. The defects are highlighted with red rectangles, and the green rectangles indicate corresponding InDiTE r…
Figure 10
Figure 10. Figure 10: Illustration of the ground-truth, the target and reference time-variant images for InDiTE boosting experiments [PITH_FULL_IMAGE:figures/full_fig_p009_10.png]
Figure 11
Figure 11. Figure 11: Visualization of inpainting results for the time-variant images in [PITH_FULL_IMAGE:figures/full_fig_p009_11.png]
Figure 12
Figure 12. Figure 12: Visualization of InDiTE and InTiDE-Diff results for three exemplar time-variant image pairs. The confidence masks are inversed for better visualization. InDiTE boosted version under tvRefInpaint and tvDuoInpaint setup. It is clear that our InDiTE generally boosted exi…
Figure 13
Figure 13. Figure 13: Visualization of another two more LeftRefill inpainting results under tvDuoInpaint setup for intuitive experiments. INTUITIVE EXPERIMENT DETAILS During the intuitive experiments, the official released code for LeftRefill 4 is utilized. We follow the official instructi…
Figure 14
Figure 14. Figure 14: Illustration of SOTA reference-guided image inpainting methods and our InDiTE for inpainting target duo-images with small (i.e., 0%-20%) mask ratio. BORDER IMPACT Our research introduces a novel approach to inpainting temporally misaligned images of the same scene, of…

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

48 extracted references · 42 canonical work pages

  1. [1]

    Repaint: Inpainting using denoising diffusion probabilistic models,

    A. Lugmayr, M. Danelljan, A. Romero, F. Yu, R. Timofte, and L. Van Gool, “Repaint: Inpainting using denoising diffusion probabilistic models,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2022, pp. 11 461–11 471. 1, 3

    work page 2022

  2. [2]

    Towards coherent image inpainting using denoising diffusion implicit models,

    G. Zhang, J. Ji, Y . Zhang, M. Yu, T. Jaakkola, and S. Chang, “Towards coherent image inpainting using denoising diffusion implicit models,” inInternational Conference on Machine Learning. PMLR, 2023, pp. 41 164–41 193. 1, 3, 7

    work page 2023

  3. [3]

    Reference-based multi-stage progressive restoration for multi-degraded images,

    Y . Zhang, Q. Yang, D. M. Chandler, and X. Mou, “Reference-based multi-stage progressive restoration for multi-degraded images,”IEEE Transactions on Image Processing, 2024. 1

    work page 2024

  4. [4]

    Dynamic selection network for image inpainting,

    N. Wang, Y . Zhang, and L. Zhang, “Dynamic selection network for image inpainting,”IEEE Transactions on Image Processing, vol. 30, pp. 1784–1798, 2021. 1, 3

    work page 2021

  5. [5]

    Jpgnet: Joint predictive filtering and generative network for image inpainting,

    Q. Guo, X. Li, F. Juefei-Xu, H. Yu, Y . Liu, and S. Wang, “Jpgnet: Joint predictive filtering and generative network for image inpainting,” in Proceedings of the 29th ACM International Conference on Multimedia, 2021, pp. 386–394. 1

    work page 2021

  6. [6]

    Misf: Multi- level interactive siamese filtering for high-fidelity image inpainting,

    X. Li, Q. Guo, D. Lin, P. Li, W. Feng, and S. Wang, “Misf: Multi- level interactive siamese filtering for high-fidelity image inpainting,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2022, pp. 1869–1878. 1, 2, 5, 7, 8

    work page 2022

  7. [7]

    Coarse-and fine-grained fusion hierarchical network for hole filling in view synthesis,

    G. Wang, K. Jiang, K. Gu, H. Liu, H. Liu, and W. Zhang, “Coarse-and fine-grained fusion hierarchical network for hole filling in view synthesis,” IEEE Transactions on Image Processing, vol. 33, pp. 322–337, 2023. 1

    work page 2023

  8. [8]

    Onion-peel networks for deep video completion,

    S. W. Oh, S. Lee, J.-Y . Lee, and S. J. Kim, “Onion-peel networks for deep video completion,” inProceedings of the IEEE/CVF international conference on computer vision, 2019, pp. 4403–4412. 1

    work page 2019

Show all 48 references
  1. [9]

    Transfill: Reference-guided image inpainting by merging multiple color and spatial transformations,

    Y . Zhou, C. Barnes, E. Shechtman, and S. Amirghodsi, “Transfill: Reference-guided image inpainting by merging multiple color and spatial transformations,” inProceedings of the IEEE/CVF CVPR, 2021, pp. 2266–2276. 1, 3, 6, 7, 8

    2021

  2. [10]

    Rego: Reference-guided outpainting for scenery image,

    Y . Wang, Y . Wei, X. Qian, L. Zhu, and Y . Yang, “Rego: Reference-guided outpainting for scenery image,”IEEE Transactions on Image Processing, vol. 33, pp. 1375–1388, 2024. 1

    2024

  3. [11]

    Transref: Multi-scale reference embedding transformer for reference- guided image inpainting,

    L. Liao, T. Liu, D. Chen, J. Xiao, Z. Wang, C.-W. Lin, and S. Satoh, “Transref: Multi-scale reference embedding transformer for reference- guided image inpainting,”arXiv:2306.11528, 2023. 1, 3, 6, 7

    2023 arXiv

  4. [12]

    Leftrefill: Filling right canvas based on left reference through generalized text-to-image diffusion model,

    C. Cao, Y . Cai, Q. Dong, Y . Wang, and Y . Fu, “Leftrefill: Filling right canvas based on left reference through generalized text-to-image diffusion model,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2024, pp. 7705–7715. 1, 2, 3, 4, 6, 7

    2024

  5. [13]

    Zero-shot image restoration using denoising diffusion null-space model,

    Y . Wang, J. Yu, and J. Zhang, “Zero-shot image restoration using denoising diffusion null-space model,”arXiv preprint arXiv:2212.00490,

    arXiv

  6. [14]

    Deep learning- based image and video inpainting: A survey,

    W. Quan, J. Chen, Y . Liu, D.-M. Yan, and P. Wonka, “Deep learning- based image and video inpainting: A survey,”International Journal of Computer Vision, pp. 1–34, 2024. 2

    2024

  7. [15]

    Image inpainting by end-to-end cascaded refinement with mask awareness,

    M. Zhu, D. He, X. Li, C. Li, F. Li, X. Liu, E. Ding, and Z. Zhang, “Image inpainting by end-to-end cascaded refinement with mask awareness,” IEEE Transactions on Image Processing, vol. 30, pp. 4855–4866, 2021. 3

    2021

  8. [16]

    Parallel multi-resolution fusion network for image inpainting,

    W. Wang, J. Zhang, L. Niu, H. Ling, X. Yang, and L. Zhang, “Parallel multi-resolution fusion network for image inpainting,” inProceedings of the IEEE/CVF international conference on computer vision, 2021, pp. 14 559–14 568. 3

    2021

  9. [17]

    W-net: Structure and texture interaction for image inpainting,

    R. Zhang, W. Quan, Y . Zhang, J. Wang, and D.-M. Yan, “W-net: Structure and texture interaction for image inpainting,”IEEE Transactions on Multimedia, 2022. 3

    2022

  10. [18]

    Masked autoencoders are scalable vision learners,

    K. He, X. Chen, S. Xie, Y . Li, P. Dollár, and R. Girshick, “Masked autoencoders are scalable vision learners,” inProceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2022, pp. 16 000–16 009. 3 JOURNAL OF LATEX CLASS FILES, VOL. 14, NO. 8, AUGUST 2021 12

    2022

  11. [19]

    Bridging global context interactions for high-fidelity image completion,

    C. Zheng, T.-J. Cham, J. Cai, and D. Phung, “Bridging global context interactions for high-fidelity image completion,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2022, pp. 11 512–11 522. 3

    2022

  12. [20]

    Aggregated contextual transfor- mations for high-resolution image inpainting,

    Y . Zeng, J. Fu, H. Chao, and B. Guo, “Aggregated contextual transfor- mations for high-resolution image inpainting,”IEEE Transactions on Visualization and Computer Graphics, 2022. 3

    2022

  13. [21]

    Glama: Joint spatial and frequency loss for general image inpainting,

    Z. Lu, J. Jiang, J. Huang, G. Wu, and X. Liu, “Glama: Joint spatial and frequency loss for general image inpainting,” inProceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2022, pp. 1301–1310. 3

    2022

  14. [22]

    Generative memory-guided semantic reasoning model for image inpainting,

    X. Feng, W. Pei, F. Li, F. Chen, D. Zhang, and G. Lu, “Generative memory-guided semantic reasoning model for image inpainting,”IEEE Transactions on Circuits and Systems for Video Technology, vol. 32, no. 11, pp. 7432–7447, 2022. 3

    2022

  15. [23]

    Dual-path image inpainting with auxiliary gan inversion,

    W. Wang, L. Niu, J. Zhang, X. Yang, and L. Zhang, “Dual-path image inpainting with auxiliary gan inversion,” inProceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2022, pp. 11 421– 11 430. 3

    2022

  16. [24]

    Zoom-to-inpaint: Image inpainting with high-frequency details,

    S. Y . Kim, K. Aberman, N. Kanazawa, R. Garg, N. Wadhwa, H. Chang, N. Karnad, M. Kim, and O. Liba, “Zoom-to-inpaint: Image inpainting with high-frequency details,” inProceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2022, pp. 477–487. 3

    2022

  17. [25]

    Image inpainting using frequency-domain priors,

    H. Roy, S. Chaudhury, T. Yamasaki, and T. Hashimoto, “Image inpainting using frequency-domain priors,”Journal of Electronic Imaging, vol. 30, no. 2, pp. 023 016–023 016, 2021. 3

    2021

  18. [26]

    Boundary-aware image inpainting with multiple auxiliary cues,

    Y . Yamashita, K. Shimosato, and N. Ukita, “Boundary-aware image inpainting with multiple auxiliary cues,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2022, pp. 619–

    2022

  19. [27]

    Deep generative model for image inpainting with local binary pattern learning and spatial attention,

    H. Wu, J. Zhou, and Y . Li, “Deep generative model for image inpainting with local binary pattern learning and spatial attention,”IEEE Transactions on Multimedia, vol. 24, pp. 4016–4027, 2021. 3

    2021

  20. [28]

    High- resolution image inpainting with iterative confidence feedback and guided upsampling,

    Y . Zeng, Z. Lin, J. Yang, J. Zhang, E. Shechtman, and H. Lu, “High- resolution image inpainting with iterative confidence feedback and guided upsampling,” inComputer Vision–ECCV 2020: 16th European Conference, Glasgow, UK, August 23–28, 2020, Proceedings, Part XIX

    2020

  21. [29]

    Springer, 2020, pp. 1–17. 3

    2020

  22. [30]

    Srinpaintor: When super-resolution meets transformer for image inpainting,

    F. Li, A. Li, J. Qin, H. Bai, W. Lin, R. Cong, and Y . Zhao, “Srinpaintor: When super-resolution meets transformer for image inpainting,”IEEE Transactions on Computational Imaging, vol. 8, pp. 743–758, 2022. 3

    2022

  23. [31]

    Auto-encoding variational bayes,

    D. P. Kingma and M. Welling, “Auto-encoding variational bayes,”arXiv preprint arXiv:1312.6114, 2013. 3

    2013 arXiv

  24. [32]

    Generative adversarial nets,

    I. Goodfellow, J. Pouget-Abadie, M. Mirza, B. Xu, D. Warde-Farley, S. Ozair, A. Courville, and Y . Bengio, “Generative adversarial nets,” Advances in neural information processing systems, vol. 27, 2014. 3

    2014

  25. [33]

    Pluralistic free-form image completion,

    C. Zheng, T.-J. Cham, and J. Cai, “Pluralistic free-form image completion,” International Journal of Computer Vision, vol. 129, no. 10, pp. 2786– 2805, 2021. 3

    2021

  26. [34]

    Image inpainting with cascaded modulation gan and object-aware training,

    H. Zheng, Z. Lin, J. Lu, S. Cohen, E. Shechtman, C. Barnes, J. Zhang, N. Xu, S. Amirghodsi, and J. Luo, “Image inpainting with cascaded modulation gan and object-aware training,” inEuropean Conference on Computer Vision. Springer, 2022, pp. 277–296. 3

    2022

  27. [35]

    Diverse image inpainting with normalizing flow,

    C. Wang, Y . Zhu, and C. Yuan, “Diverse image inpainting with normalizing flow,” inEuropean conference on computer vision. Springer, 2022, pp. 53–69. 3

    2022

  28. [36]

    High-fidelity pluralistic image completion with transformers,

    Z. Wan, J. Zhang, D. Chen, and J. Liao, “High-fidelity pluralistic image completion with transformers,” inProceedings of the IEEE/CVF International Conference on Computer Vision, 2021, pp. 4692–4701. 3

    2021

  29. [37]

    Denoising diffusion probabilistic models,

    J. Ho, A. Jain, and P. Abbeel, “Denoising diffusion probabilistic models,” Advances in neural information processing systems, vol. 33, pp. 6840– 6851, 2020. 3

    2020

  30. [38]

    Reference-guided texture and structure inference for image inpainting,

    T. Liu, L. Liao, Z. Wang, and S. Satoh, “Reference-guided texture and structure inference for image inpainting,” in2022 IEEE international conference on image processing (ICIP). IEEE, 2022, pp. 1996–2000. 3

    2022

  31. [39]

    Reference-guided landmark image inpainting with deep feature matching,

    J. Li, Z. Xiong, and D. Liu, “Reference-guided landmark image inpainting with deep feature matching,”IEEE Transactions on Circuits and Systems for Video Technology, vol. 32, no. 12, pp. 8422–8435, 2022. 3

    2022

  32. [40]

    Learning dynamic siamese network for visual object tracking,

    Q. Guo, W. Feng, C. Zhou, R. Huang, L. Wan, and S. Wang, “Learning dynamic siamese network for visual object tracking,” inProceedings of the IEEE international conference on computer vision, 2017, pp. 1763–1771. 5

    2017

  33. [41]

    U-net: Convolutional networks for biomedical image segmentation,

    O. Ronneberger, P. Fischer, and T. Brox, “U-net: Convolutional networks for biomedical image segmentation,” inMedical image computing and computer-assisted intervention–MICCAI 2015: 18th international conference, Munich, Germany, October 5-9, 2015, proceedings, part III

    2015

  34. [42]

    Springer, 2015, pp. 234–241. 5

    2015

  35. [43]

    Back to the source: Diffusion-driven test-time adaptation,

    J. Gao, J. Zhang, X. Liu, T. Darrell, E. Shelhamer, and D. Wang, “Back to the source: Diffusion-driven test-time adaptation,”arXiv preprint arXiv:2207.03442, 2022. 6

    2022 arXiv

  36. [44]

    Edgeconnect: Structure guided image inpainting using edge prediction,

    K. Nazeri, E. Ng, T. Joseph, F. Qureshi, and M. Ebrahimi, “Edgeconnect: Structure guided image inpainting using edge prediction,” inProceedings of the IEEE/CVF international conference on computer vision workshops, 2019, pp. 0–0. 6, 8

    2019

  37. [45]

    Ilvr: Conditioning method for denoising diffusion probabilistic models,

    J. Choi, S. Kim, Y . Jeong, Y . Gwon, and S. Yoon, “Ilvr: Conditioning method for denoising diffusion probabilistic models,”arXiv preprint arXiv:2108.02938, 2021. 6

    2021 arXiv

  38. [46]

    Street- view change detection with deconvolutional networks,

    P. F. Alcantarilla, S. Stent, G. Ros, R. Arroyo, and R. Gherardi, “Street- view change detection with deconvolutional networks,”Autonomous Robots, vol. 42, pp. 1301–1322, 2018. 7

    2018

  39. [47]

    Image inpainting for irregular holes using partial convolutions,

    G. Liu, F. A. Reda, K. J. Shih, T.-C. Wang, A. Tao, and B. Catanzaro, “Image inpainting for irregular holes using partial convolutions,” in Proceedings of the European conference on computer vision (ECCV), 2018, pp. 85–100. 7

    2018

  40. [48]

    Benchmarking neural network ro- bustness to common corruptions and perturbations,

    D. Hendrycks and T. Dietterich, “Benchmarking neural network ro- bustness to common corruptions and perturbations,”arXiv preprint arXiv:1903.12261, 2019. 13 JOURNAL OF LATEX CLASS FILES, VOL. 14, NO. 8, AUGUST 2021 13 Fig. 13:Visualization of another two more LeftRefill inpain...

    1903 arXiv

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.