Organizations: Beijing University of Posts and Telecommunications · Department of Computer Science and Technology, Institute for AI, Tsinghua University · Minzu University of China · Xiaomi Corporation
To better preserve an individual's identity, face restoration has evolved from reference-free to reference-based approaches, which leverage high-quality reference images of the same identity to enhance identity fidelity in the restored outputs. However, most existing methods implicitly assume that the reference and degraded input are age-aligned, limiting their effectiveness in real-world scenarios where only cross-age references are available, such as historical photo restoration. This paper proposes MeInTime, a diffusion-based face restoration method that extends reference-based restoration from same-age to cross-age settings. Given one or few reference images along with an age prompt corresponding to the degraded input, MeInTime achieves faithful restoration with both identity fidelity and age consistency. Specifically, we decouple the modeling of identity and age conditions. During training, we focus solely on effectively injecting identity features through a newly introduced attention mechanism and introduce Gated Residual Fusion modules to facilitate the integration between degraded features and identity representations. At inference, we propose Age-Aware Gradient Guidance, a training-free sampling strategy, using an age-driven direction to iteratively nudge the identity-aware denoising latent toward the desired age semantic manifold. Extensive experiments demonstrate that MeInTime outperforms existing face restoration methods in both identity preservation and age consistency. Our code is available at: https://github.com/teer4/MeInTime
Figures & tables
Figure 1: Overview of MeInTime. (a) During training, identity features from reference images are extracted by Insightface ( Deng et al., 2019 ) , projected into face embeddings, and injected into UNet via decoupled cross-attention with a unified prompt “photo of a person”. Gated Residual Fusion (GRF) modules are introduced into each decoder block to facilitate feature fusion. The degraded input is processed following DiffBIR, which incorporates a restoration model ( Liang et al., 2021 ) and the proposed IRControlNet as structural guidance. We adopt the standard diffusion loss ( LDiff ) as the training objective. (b) During inference, given the target age, the framework generates an age prompt ( e.g. , “photo of a 24-year-old person”) and performs two forward passes–with and without the age condition–to compute an age-aware gradient that iteratively refines the latent along the denoising process, enabling identity-preserving, age-controllable restoration.
Figure 2: The structure of Gated Residual Fusion module.
Figure 3: Visual comparison of identity-preserving restoration under different age control strategies.
Figure 4: Visual comparison of different optimization steps.
Same-Age
Cross-Age
Method
Ref
PSNR ↑
SSIM ↑
LPIPS ↓
MUSIQ ↑
FID ↓
IDS ↑
PSNR ↑
SSIM ↑
LPIPS ↓
MUSIQ ↑
FID ↓
IDS ↑
AGE ↓
CodeFormer
25.61
0.720
0.207
75.87
47.46
0.639
26.29
0.749
0.142
75.77
55.85
0.519
11.13
DR2+SPAR
21.97
0.681
0.312
71.31
69.43
0.213
21.71
0.646
0.357
64.52
95.05
0.118
15.52
DifFace
25.31
0.720
0.247
70.00
55.01
0.509
24.94
0.691
0.274
65.31
83.73
0.410
11.59
DMDNet
✓
25.76
0.730
0.228
73.92
55.48
0.703
25.76
0.712
0.233
72.94
66.37
0.621
13.27
RestorerID
✓
25.00
0.699
0.272
72.75
59.74
0.686
24.94
0.679
0.242
75.71
69.18
0.618
15.37
Table 1: Quantitative comparison with state-of-the-art methods on same-age and cross-age face restoration. The best results are shown in yellow , and the second-best are blue .
Figure 5: User study results.
Figure 6: Identity Preservation vs. Age Consistency.
Figure 7: Comparison with state-of-the-art methods on same-age data. Zoom in for best view.
Figure 8: Comparison with state-of-the-art methods on cross-age data. Zoom in for best view.
Figure 10Figure 11
Appendix figures & tables12 assets
Supplementary material from the paper’s appendix.
Appendix
Figure 12: Statistics of cross-age identity datasets.
Figure 13: Cross-age testing data instance.
Figure 14: Visualization of identity-token attention maps. Each column shows one restoration run with a different-age reference of the same identity. The second row gives the restored results, and the bottom row shows the averaged 16×16 identity-token attention maps across all steps.
Method
Ref
MUSIQ ↑
FID ↓
IDS ↑
AGE ↓
CodeFormer
70.17
56.53
0.263
9.67
DR2+SPAR
65.09
123.50
0.229
10.23
DifFace
56.28
119.78
0.295
11.83
DMDNet
✓
69.79
105.54
0.295
9.67
RestorerID
✓
74.25
76.31
0.467
15.14
Ref-LDM
✓
68.04
59.93
0.352
13.33
Appendix
Table 5: Quantitative comparison on the real-world dataset. The best results are shown in yellow , and the second-best are blue .
Figure 15: Comparison with state-of-the-art methods on real-world data. Zoom in for best view.
Method
Type
PSNR ↑
LPIPS ↓
MUSIQ ↑
IDS ↑
AGE ↓
Time(s)
HRFAE
GAN
21.57
0.308
67.31
0.602
10.23
0.18
SAM
GAN
18.09
0.457
60.33
0.347
8.77
0.45
FADING
Diffusion
22.28
0.270
66.52
0.584
8.05
133.41
MeInTime
Diffusion
24.53
0.208
74.33
0.655
6.89
43.37
Appendix
Table 6: Quantitative assessment of post-editing effects on FaceMe-restored images.
Figure 16: Post-hoc age editing applied to restored images from RestorerID, Ref-LDM, and FaceMe. Existing age-editing models either under-edit, over-edit, or damage identity, demonstrating the limitation of two-stage pipelines for cross-age restoration.
Methods
DifFace
DR2
RestorerID
Ref-LDM
FaceMe
MeInTime (N=0)
MeInTime (N=5)
Time(s)
5.96
1.23
7.74
1.79
7.65
7.26
43.37
Appendix
Table 7: Inference time comparison.
Figure 17: Over-sharpening artifacts when conditioning on high-age prompts.
Figure 18: Comparison with state-of-the-art methods on same-age data. Zoom in for best view.
Figure 19: Comparison with state-of-the-art methods on cross-age data. Zoom in for best view.
Figure 20: Example set used in the user study. For each question, participants are shown the ground-truth image, a reference image, and multiple restored results (A–E). They select the option that best satisfies the evaluation criterion for that question. The target age indicated at the top ( e.g. , “30-year-old”) is used only for age-related questions.
Recent progress in face restoration has shifted from visual fidelity to identity fidelity, driving a transition from reference-free to reference-based paradigms that condition restoration on reference images of the same person. However, these methods assume the reference and degraded input are age-aligned. When only cross-age references are available, as in historical restoration or missing-person retrieval, they fail to maintain age fidelity. To address this limitation, we propose TimeWeaver, the first reference-based face restoration framework supporting cross-age references. Given arbitrary reference images and a target-age prompt, TimeWeaver produces restorations with both identity fidelity and age consistency. Specifically, we decouple identity and age conditioning across training and inference. During training, the model learns an age-robust identity representation by fusing a global identity embedding with age-suppressed facial tokens via a transformer-based ID-Fusion module. During inference, two training-free techniques, Age-Aware Gradient Guidance and Token-Targeted Attention Boost, steer sampling toward desired age semantics, enabling precise adherence to the target-age prompt. Extensive experiments show that TimeWeaver surpasses existing methods in visual quality, identity preservation, and age consistency.
Teer Song, Yue Zhang, Yu Tian +6
Beijing University of Posts and Telecommunications · Tsinghua University · Minzu University of China +1
Blind face restoration is highly ill-posed under severe degradation, where identity-critical details may be missing from the degraded input. Same-identity references reduce this ambiguity, but mismatched pose, expression, illumination, age, makeup, or local facial states can lead to overuse of reference appearance. We propose IConFace, a unified reference-aware and no-reference framework with identity--structure asymmetric conditioning. References are distilled into a norm-weighted global AdaFace identity anchor for image-only modulation, while the degraded image is reinforced as the spatial structure anchor through low-rank residuals and block-wise degraded cross-attention with two-route memory. The resulting single checkpoint exploits references when available and falls back to no-reference restoration when absent, improving identity consistency, fine-detail recovery, and degraded-only restoration quality in a unified model.
Face aging, an ill-posed problem shaped by environmental and genetic factors, is vital in entertainment, forensics, and digital archiving, where realistic age transformations must preserve both identity and visual realism. However, existing works relying on numerical age representations overlook the interplay of biological and contextual cues. Despite progress in recent face aging models, they struggle with identity preservation in wide age transformations, also static attention and optimization-heavy inversion in diffusion limit adaptability, fine-grained control and background consistency. To address these challenges, we propose Face Time Traveller (FaceTT), a diffusion-based framework that achieves high-fidelity, identity-consistent age transformation. Here, we introduce a Face-Attribute-Aware Prompt Refinement strategy that encodes intrinsic (biological) and extrinsic (environmental) aging cues for context-aware conditioning. A tuning-free Angular Inversion method is proposed that efficiently maps real faces into the diffusion latent space for fast and accurate reconstruction. Moreover, an Adaptive Attention Control mechanism is introduced that dynamically balances cross-attention for semantic aging cues and self-attention for structural and identity preservation. Extensive experiments on benchmark datasets and in-the-wild testset demonstrate that FaceTT achieves superior identity retention, background preservation and aging realism over state-of-the-art (SOTA) methods.