Pith. sign in

REVIEW 2 major objections 5 minor 100 references

PersonaCraft: Personalized and Controllable Full-Body Multi-Human Scene Generation Using Occlusion-Aware 3D-Conditioned Diffusion

T0 review · 2 major / 5 minor · reviewed 2026-08-12 · deepseek-v4-flash

Pith's one-line read PersonaCraft uses SMPLx-derived 3D geometry—depth and normal maps—to condition a diffusion model for multi-human image generation, reporting large gains in pose accuracy, face identity, and body-shape preservation over 2D-skeleton…

desk verdict The system is a real multi-person 3D-conditioned generation contribution, but the headline body-shape metric is self-referential and the pose comparisons are asymmetric; deserves a serious referee, conditional on an independent body-shape evaluation and code/data release. read the letter →

arxiv 2411.18068 v2 pith:AWSGXL3K submitted 2024-11-27 cs.CV cs.AI

classification cs.CVcs.AI
keywords diffusionmodelscontrollableimagegenerationSMPLx3Dhumanposeconditioningocclusionhandlingfull-bodypersonalizationmulti-humansceneclassifier-freeguidance
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

PersonaCraft's central claim is that conditioning a diffusion image generator on 3D human-body geometry, rather than on 2D skeletons, removes the ambiguity behind occlusion failures and anatomical distortions in multi-person scenes. The paper builds SMPLx mesh renderings (depth and normal maps) into a ControlNet-style branch, adds an Occlusion Boundary Enhancer Network that trains on depth edges inside occlusion masks, and applies occlusion-aware classifier-free guidance so conditioning is strengthened exactly where bodies overlap. It combines these with a face-identity branch and a dual-pathway body-shape representation to personalize both face and body. In their evaluations, the method reports substantially lower 3D pose error (MPJPE 60.65 vs. 112.6 for the strongest personalized baseline) and higher body-shape preservation, and user studies rank it first for naturalness, identity, and pose consistency.

What carries the argument

The machinery has three cooperating parts. SMPLx-ControlNet (SCNet) feeds SMPLx depth and normal renderings through a ControlNet, giving the U-Net depth cues and surface orientation; depth alone is best among single renderings, and depth-plus-normal is the chosen default. The Occlusion Boundary Enhancer Network (OccNet) receives only SMPLx depth edges inside occlusion masks, forcing the model to learn boundary structure where bodies overlap. Occlusion-Aware Classifier-Free Guidance (OccCFG) then scales the conditional noise prediction by a larger factor inside the occlusion mask and a smaller factor outside, so occluded regions get strengthened conditioning without over-saturating open areas. Full-body personalization is carried by SMPLx shape coefficients in the rendered geometry plus an optional CLIP-based textual body-shape pathway, with face identity injected through a face ControlNet.

What would settle it

Take a set of heavily occluded multi-person images for which SMPLx fits are known to be poor, or compare against motion-capture ground-truth meshes in the same poses. If PersonaCraft with those poor fits does not outperform a 2D-skeleton baseline, or if its MPJPE advantage disappears when the fitter is held fixed across methods, then the occlusion robustness is coming from the fitter rather than from the diffusion-side machinery.

Watch

Extended reading notes

Core claim

The discovery the paper tries to establish is that a single diffusion pipeline can jointly solve full-body identity preservation and occlusion-heavy multi-person generation when pose is represented by SMPLx depth and normal maps. The authors show that 3D conditioning alone still misses fine-grained occlusions, so they add occlusion masks (pixels whose ray intersects more than two surfaces), an Occlusion Boundary Enhancer Network that learns from depth-edge signals inside those masks, and an occlusion-aware classifier-free guidance rule that applies a higher CFG scale only in occluded regions. Combined with a face identity ControlNet and a body-shape pathway that uses SMPLx shape coefficients plus optional textual refinement, the system claims personalized and controllable multi-human synthesis. Reported numbers include MPJPE 62.647 vs. 102.64 for pose-controlled baselines and top-1 user preference above 40% across all evaluated criteria.

Load-bearing premise

The load-bearing premise is that the SMPLx pose and shape parameters recovered from ordinary RGB photos are accurate enough to serve as both the generation condition and the evaluation ground truth; if those fits are wrong in crowded or occluded scenes, the whole pipeline and its metrics inherit the error.

Editorial extensions

If this is right

  • Replacing 2D skeleton conditioning with SMPLx depth and normal conditioning is enough to reduce 3D pose error in multi-person scenes by roughly half in their tests.
  • Occlusion masks derived from 3D mesh intersections let the model locate fine occlusions, and depth edges inside those masks are a sufficient training signal to sharpen occluded anatomy.
  • Raising classifier-free guidance inside occluded regions only, rather than globally, improves anatomical consistency without the usual over-saturation.
  • Full-body personalization can be built by separating body shape (SMPLx coefficients plus text) from face identity, which existing face-only personalizers do not do.
  • Because the conditioning is modular, the same SCNet and OccNet residual additions can be combined with different face modules and style adapters without retraining the whole pipeline.

Reading between the lines

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

  • If the reported MPJPE gains largely reflect the single-shot SMPLx fitter's own estimates acting as both conditioning and evaluation ground truth, then a fair comparison against methods using motion-capture or manually registered 3D poses would separate the value of the 3D representation from the value of the fitter; the paper does not run that experiment.
  • The occlusion-mask idea is not human-specific: counting intersected surfaces and masking depth edges could be applied to any articulated object in a ControlNet-style diffusion model, though the paper does not test this.
  • Because OccCFG only needs an occlusion mask at inference, it could be plugged into other 3D-conditioned human generators; a direct test would be running it on a single-person generator with self-occlusion and checking whether the mask-based guidance alone reproduces the reported gains.
  • A likely failure mode the paper acknowledges is that poor SMPLx fits in partially visible references will propagate into both generation and evaluation, so a user-facing system would need a rejection rule for low-confidence fits; the paper does not propose one.
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

2 major / 5 minor

Summary. PersonaCraft proposes a diffusion-based framework for multi-human personalized image synthesis. It conditions an SDXL generator on SMPLx depth and normal maps through a fine-tuned ControlNet (SCNet), adds an Occlusion Boundary Enhancer Network (OccNet) trained on masked depth edges, and applies Occlusion-Aware Classifier-Free Guidance (OccCFG) to reinforce conditioning in occluded regions. Full-body personalization is achieved by injecting SMPLx shape parameters through SCNet and combining them with Face Identity ControlNet, plus an optional textual body-shape refinement pathway. The paper reports extensive experiments on COCO-WholeBody, with quantitative and user-study results that claim large improvements over baselines in face identity, body shape, pose accuracy, and perceptual quality.

Significance. The architectural components are well motivated, and the ablations in Tab. 5 show that each proposed module contributes to pose consistency. If the body-shape preservation metric is replaced with an independent evaluation, the method would represent a substantial advance in controllable multi-human generation with explicit occlusion handling and full-body personalization. The paper also provides a useful efficiency analysis and a broad qualitative comparison, and it ships a supplementary description of training and evaluation details.

major comments (2)
  1. [Supp. Eq. (S7), Sec. 3.3, Tab. 1] The body-shape preservation score S_body in Supp. Eq. (S7) is self-referential: β_ref is estimated by MultiHMR from the reference image and is exactly the vector rendered into SMPLx depth/normal maps used as conditioning in Sec. 3.3, while β_gen is also obtained by running MultiHMR on the generated image. The score therefore measures consistency with the conditioning estimator, not independently verified anatomical fidelity, and the comparison in Tab. 1 is asymmetric because the OMG/IPAdapter baselines do not receive β_ref as conditioning. An independent body-shape regressor (a different SMPL/SMPLx fitter or a learned shape descriptor) should be used for both reference and generated images, or the claim should be restricted to conditioning fidelity. The user study in Tab. 2 is suggestive, but the protocol in Sec. B.4 does not report identity-masked, randomized presentation, so it does not currently rule out the metric artifact.
  2. [Eq. (5), Eq. (3), Sec. 3.2, Sec. 3.3] The inference-time application of OccNet appears inconsistent with its training objective. Eq. (3) trains OccNet on e_SMPLx ⊙ M_occ (depth edges inside the occlusion mask), and the text of Sec. 3.2 states that the model is provided only these edges. However, Eq. (5) defines R_occ = E^Occ_ω(d_SMPLx), i.e., the full depth map rather than the edge signal. If this is a typo, it should be corrected to e_SMPLx ⊙ M_occ; if not, the discrepancy between training and inference inputs must be explained, since the ablation in Tab. 5 attributes a pose-consistency gain to OccNet.
minor comments (5)
  1. [Supp. Tab. S6] The row 'w/ body shape' reports Total = 0.615, but Tab. 1 reports Total = 0.560 for the same Single (0.630) and Multi (0.548) values; this numerical inconsistency should be corrected or explained.
  2. [Sec. 4, Metrics] The MPJPE and AP-0.5 metrics are computed by running MultiHMR on generated images while the target poses also come from MultiHMR; this is acceptable as a relative comparison, but the paper should state that these scores measure compatibility with MultiHMR's pose space rather than absolute 3D accuracy.
  3. [Sec. B.4, User Study] The user study description should state whether participants were blinded to method identity and whether the order of presented images was randomized; without this information, the preference rates in Tabs. 2 and 4 are difficult to interpret.
  4. [Sec. 4.2, Baselines] The baseline 'ControlNet-Flux' is cited as [48, 91]; it would be cleaner to cite the Flux model alone and note that ControlNet is applied on top, since [48] is the base model and [91] is the ControlNet method.
  5. [Sec. 3.3, Dual-Pathway Body Shape Personalization] The textual body-shape pathway is only evaluated qualitatively (Fig. 11); a quantitative ablation of this pathway, even on a small set, would strengthen the claim of dual-pathway benefit.

Circularity Check

1 steps flagged · score 6.0 of 10

The headline body-shape preservation score in Tab. 1 is computed in the same MultiHMR-vs-MultiHMR beta-space that is fed into SCNet as conditioning, so the quantitative claim reduces largely to a self-consistency check rather than an independent test of anatomical fidelity.

  1. fitted input called prediction [Sec. 3.3 (Full-Body Personalized Generation), Sec. 4 (Metrics), Supplementary B.2 Eq. (S7)]
    "To achieve this, we use MultiHMR [12] to estimate SMPLx body shape parameters beta(i) ... SMPLx renderings, like depth dSMPLx, serve as conditioning signals for the diffusion model ... Body shape preservation was evaluated using cosine similarity between the SMPLx body shape parameters beta from the reference and generated instances ... Sbody = ... beta_ref,i · beta_gen,i / (||beta_ref,i|| ||beta_gen,i||)"

    The reference shape beta_ref is not an external ground truth: it is estimated by MultiHMR from the reference image and rendered into the SMPLx depth/normal maps that condition SCNet. The reported metric Sbody then re-runs MultiHMR on the generated image and measures cosine similarity to the very same beta_ref. The generator is trained and conditioned to be consistent with MultiHMR's beta-space, so a high score can be achieved by producing images that MultiHMR reads back as the conditioning vector, whether or not the true body shape is accurate. Baselines conditioned only on 2D pose or face embeddings do not receive beta_ref, so the Tab. 1 margin (0.560 vs 0.401) partly reflects this asymmetry rather than anatomical fidelity. The Sec.

full rationale

The paper presents a controllable image-generation system rather than a formal derivation, so most of its pipeline is self-contained and not circular. The one clear circular step is the body-shape preservation metric: beta is both the conditioning input (estimated by MultiHMR, rendered into SCNet depth/normal maps) and the measurement target (estimated again by MultiHMR from the generated image, compared via Eq. S7). This makes the Tab. 1 body-shape score a conditioning-fidelity check in MultiHMR's beta-space, not an independent verification of body-shape accuracy. The face-identity metric is less problematic because conditioning uses InsightFace embeddings while evaluation uses FaceNet, so the two networks are not identical. The pose metrics are also computed relative to the SMPLx/MultiHMR conditioning space, but the paper does not explicitly state which estimator is applied to generated images, so I do not count that as a separate demonstrated step. The user studies provide independent perceptual evidence, and no load-bearing self-citation chain was found, so the partial circularity is limited to the quantitative body-shape and closely related conditioning-fidelity metrics. Overall score 6 reflects that one headline quantitative prediction reduces by construction to its own conditioning input, while the central claim retains independent perceptual support.

Assumptions & free parameters 5 free parameters · 4 assumptions · 0 invented entities

The central claim rests on an external 3D fitting tool (MultiHMR) assumed accurate enough for conditioning and evaluation, on hand-set thresholds defining occlusions (Eq. 2), and on fine-tuned CFG and conditioning weights chosen by experiment. No new physical or conceptual entities are introduced.

free parameters (5)
  • kbase (base CFG scale) = 3
    Set for all regions; chosen by tuning to avoid oversaturation (Sec. 4, Fig. 10).
  • kocc (occlusion-region CFG scale) = 5
    Chosen to strengthen pose conditioning in occluded regions; ablation Fig. 10.
  • Conditioning weights alpha_sc, alpha_occ, alpha_id = 0.8
    Set to 0.8 for all conditions; ablation Fig. S21 shows sensitivity.
  • MPII single-to-multi ratio = 2:1
    Introduced to balance occlusion scenarios in training set (Supp B.1); hand-chosen.
  • Occlusion threshold for number of intersected faces = >2 surfaces
    A pixel is occluded if more than two surfaces are intersected (Eq. 2); threshold chosen manually and masks are dilated.
assumptions (4)
  • domain assumption MultiHMR returns sufficiently accurate SMPLx pose and shape parameters from single RGB images of multi-person scenes, for both conditioning and evaluation.
    The entire pipeline constructs SMPLx depth/normal conditioning from MultiHMR outputs, and the body shape metric also estimates beta from generated images with MultiHMR (Sec. 3.3, Supp B.1).
  • domain assumption Face Identity ControlNet and InsightFace embeddings preserve identity when combined with SCNet and OccNet residuals.
    The framework adds IdentityNet residuals (Eq. 5) and assumes no identity leakage or loss; ablation Fig. S21 partially supports this.
  • domain assumption An occluded pixel is one where a ray intersects more than two SMPLx faces (Eq. 2), and this definition captures perceptually relevant occlusions.
    This definition is used to generate training masks and OccCFG weights; the threshold of two surfaces is hand-set and not independently validated.
  • domain assumption Pretrained SDXL and ControlNet can be fine-tuned to accept SMPLx depth, normal, and occlusion-edge renderings as conditioning without losing generation quality.
    Core training setup (Eq. 1, Eq. 3) relies on this transfer; no formal guarantee is provided.

how reviews work

0 comments
Cite this review

Pith. "Pith review of PersonaCraft: Personalized and Controllable Full-Body Multi-Human Scene Generation Using Occlusion-Aware 3D-Conditioned Diffusion." pith.science (2026). https://pith.science/paper/AWSGXL3K

@misc{pith2026241118068,
  author       = {Pith},
  title        = {Pith review of: PersonaCraft: Personalized and Controllable Full-Body Multi-Human Scene Generation Using Occlusion-Aware 3D-Conditioned Diffusion},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/AWSGXL3K}},
  note         = {Machine review of arXiv:2411.18068}
}
read the original abstract

We present PersonaCraft, a framework for controllable and occlusion-robust full-body personalized image synthesis of multiple individuals in complex scenes. Current methods struggle with occlusion-heavy scenarios and complete body personalization, as 2D pose conditioning lacks 3D geometry, often leading to ambiguous occlusions and anatomical distortions, and many approaches focus solely on facial identity. In contrast, our PersonaCraft integrates diffusion models with 3D human modeling, employing SMPLx-ControlNet, to utilize 3D geometry like depth and normal maps for robust 3D-aware pose conditioning and enhanced anatomical coherence. To handle fine-grained occlusions, we propose Occlusion Boundary Enhancer Network that exploits depth edge signals with occlusion-focused training, and Occlusion-Aware Classifier-Free Guidance strategy that selectively reinforces conditioning in occluded regions without affecting unoccluded areas. PersonaCraft can seamlessly be combined with Face Identity ControlNet, achieving full-body multi-human personalization and thus marking a significant advancement beyond prior approaches that concentrate only on facial identity. Our dual-pathway body shape representation with SMPLx-based shape parameters and textual refinement, enables precise full-body personalization and flexible user-defined body shape adjustments. Extensive quantitative experiments and user studies demonstrate that PersonaCraft significantly outperforms existing methods in generating high-quality, multi-person images with accurate personalization and robust occlusion handling.

Figures

Figures reproduced from arXiv: 2411.18068 by the authors.

Figure 1
Figure 1. PersonaCraft generates realistic, personalized images of multiple individuals with complex occlusions, preserving facial identity and body shape using occlusion-aware 3D pose and shape conditioned diffusion. PersonaCraft outperforms baselines in body shape personalization (blue arrows indicate failures) and naturalness (yellow arrows highlight artifacts, with zoomed-in views in yellow boxes). Abstract We present Per… view at source ↗
Figure 2
Figure 2. SMPLx-ControlNet (SCNet) with SMPLx depth and nor [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 3
Figure 3. Overview of our occlusion-aware 3D pose and shape conditioning. We generate SMPLx renderings for SMPLx-ControlNet ( [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗
Figures from the paper (8 more)
Figure 5
Figure 5. Figure 5: Full pipeline of PersonaCraft for multi-human full-body [PITH_FULL_IMAGE:figures/full_fig_p004_5.png]
Figure 6
Figure 6. Figure 6: Qualitative comparison of personalized multi-human scene generation. [PITH_FULL_IMAGE:figures/full_fig_p005_6.png]
Figure 7
Figure 7. Figure 7: Comparison of personalized multi-human scene generation with the methods which are not direct baselines. [PITH_FULL_IMAGE:figures/full_fig_p006_7.png]
Figure 8
Figure 8. Figure 8: Qualitative comparison of pose-controlled multi-human generation. [PITH_FULL_IMAGE:figures/full_fig_p006_8.png]
Figure 9
Figure 9. Figure 9: Effect of our occlusion-aware 3D pose & shape condition [PITH_FULL_IMAGE:figures/full_fig_p007_9.png]
Figure 10
Figure 10. Figure 10: Effect of our occlusion-aware CFG (OccCFG). Each [PITH_FULL_IMAGE:figures/full_fig_p007_10.png]
Figure 11
Figure 11. Figure 11: Effectiveness of dual-pathway body shape personaliza [PITH_FULL_IMAGE:figures/full_fig_p008_11.png]
Figure 12
Figure 12. Figure 12: Result of PersonaCraft’s user-defined body shape con [PITH_FULL_IMAGE:figures/full_fig_p008_12.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

100 extracted references · 44 canonical work pages

  1. [1]

    Github, https: //github.com/deepinsight/insightface

    Insightface: 2d and 3d face analysis project. Github, https: //github.com/deepinsight/insightface . 4, 6, 25

  2. [2]

    A neural space-time representation for text-to- image personalization

    Yuval Alaluf, Elad Richardson, Gal Metzer, and Daniel Cohen-Or. A neural space-time representation for text-to- image personalization. Transactions on Graphics, 42(6):1–10,

  3. [3]

    frosting-lane-lora-sdxl

    Alvdansen. frosting-lane-lora-sdxl. Hugging Face, https: //huggingface.co/alvdansen/frosting-lane, . 13

  4. [4]

    gemini-anime-lora-sdxl

    Alvdansen. gemini-anime-lora-sdxl. Hugging Face, https: //huggingface.co/alvdansen/geminianime , . 13

  5. [5]

    painting-light-lora-sdxl

    Alvdansen. painting-light-lora-sdxl. Hugging Face, https : / / huggingface . co / alvdansen / paintinglight/tree/main, . 13

  6. [6]

    softpastel-anime-lora-sdxl

    Alvdansen. softpastel-anime-lora-sdxl. Hugging Face, https : / / huggingface . co / alvdansen / softpastelanime, . 13

  7. [7]

    2d human pose estimation: New benchmark and state of the art analysis

    Mykhaylo Andriluka, Leonid Pishchulin, Peter Gehler, and Bernt Schiele. 2d human pose estimation: New benchmark and state of the art analysis. In CVPR, 2014. 7, 25

  8. [8]

    Moab Arar, Rinon Gal, Yuval Atzmon, Gal Chechik, Daniel Cohen-Or, Ariel Shamir, and Amit H. Bermano. Domain- agnostic tuning-encoder for fast personalization of text-to- image models. In SIGGRAPH Asia 2023, 2023. 2

Show all 100 references
  1. [9]

    Break-a-scene: Extracting multiple con- cepts from a single image

    Omri Avrahami, Kfir Aberman, Ohad Fried, Daniel Cohen-Or, and Dani Lischinski. Break-a-scene: Extracting multiple con- cepts from a single image. arXiv preprint arXiv:2305.16311,

  2. [10]

    Multidiffusion: Fusing diffusion paths for controlled image generation

    Omer Bar-Tal, Lior Yariv, Yaron Lipman, and Tali Dekel. Multidiffusion: Fusing diffusion paths for controlled image generation. arXiv preprint arXiv:2302.08113, 2023. 7, 25

  3. [11]

    Multi-hmr: Multi-person whole-body hu- man mesh recovery in a single shot

    Fabien Baradel*, Matthieu Armando, Salma Galaaoui, Ro- main Brégier, Philippe Weinzaepfel, Grégory Rogez, and Thomas Lucas*. Multi-hmr: Multi-person whole-body hu- man mesh recovery in a single shot. In ECCV, 2024. 7, 25

  4. [12]

    Multi-hmr: Multi-person whole-body human mesh recovery in a single shot

    Fabien Baradel, Matthieu Armando, Salma Galaaoui, Romain Brégier, Philippe Weinzaepfel, Grégory Rogez, and Thomas Lucas. Multi-hmr: Multi-person whole-body human mesh recovery in a single shot. In ECCV, pages 202–218. Springer,

  5. [13]

    Improving image generation with better captions

    James Betker, Gabriel Goh, Li Jing, Tim Brooks, Jianfeng Wang, Linjie Li, Long Ouyang, Juntang Zhuang, Joyce Lee, Yufei Guo, et al. Improving image generation with better captions. Computer Science. https://cdn. openai. com/papers/dall-e-3. pdf, 2(3), 2023. 2

  6. [14]

    Demystifying mmd gans

    Mikołaj Bi´nkowski, Danica J Sutherland, Michael Arbel, and Arthur Gretton. Demystifying mmd gans. arXiv, 2018. 7

  7. [15]

    graphic-novel-illustration-lora-sdxl

    Blink7630. graphic-novel-illustration-lora-sdxl. Hug- ging Face, https://huggingface.co/blink7630/ graphic-novel-illustration. 13

  8. [16]

    Con- trolling human shape and pose in text-to-image diffusion mod- 9 els via domain adaptation

    Benito Buchheim, Max Reimann, and Jürgen Döllner. Con- trolling human shape and pose in text-to-image diffusion mod- 9 els via domain adaptation. arXiv preprint arXiv:2411.04724,

  9. [17]

    Realtime multi-person 2d pose estimation using part affinity fields

    Zhe Cao, Tomas Simon, Shih-En Wei, and Yaser Sheikh. Realtime multi-person 2d pose estimation using part affinity fields. In CVPR, 2017. 3

  10. [18]

    In- structbooth: Instruction-following personalized text-to-image generation

    Daewon Chae, Nokyung Park, Jinkyu Kim, and Kimin Lee. In- structbooth: Instruction-following personalized text-to-image generation. arXiv preprint arXiv:2312.03011, 2023. 2

  11. [19]

    Subject-driven text-to-image generation via apprenticeship learning

    Wenhu Chen, Hexiang Hu, Yandong Li, Nataniel Rui, Xuhui Jia, Ming-Wei Chang, and William W Cohen. Subject-driven text-to-image generation via apprenticeship learning. arXiv preprint arXiv:2304.00186, 2023. 2

  12. [20]

    Anydoor: Zero-shot object-level im- age customization

    Xi Chen, Lianghua Huang, Yu Liu, Yujun Shen, Deli Zhao, and Hengshuang Zhao. Anydoor: Zero-shot object-level im- age customization. arXiv preprint arXiv:2307.09481, 2023. 2

  13. [21]

    Custom-edit: Text-guided image editing with cus- tomized diffusion models

    Jooyoung Choi, Yunjey Choi, Yunji Kim, Junho Kim, and Sun- groh Yoon. Custom-edit: Text-guided image editing with cus- tomized diffusion models. arXiv preprint arXiv:2305.15779,

  14. [22]

    An im- age is worth one word: Personalizing text-to-image generation using textual inversion

    Rinon Gal, Yuval Alaluf, Yuval Atzmon, Or Patashnik, Amit H Bermano, Gal Chechik, and Daniel Cohen-Or. An im- age is worth one word: Personalizing text-to-image generation using textual inversion. ICLR, 2022. 2, 7, 14, 26

  15. [23]

    Designing an encoder for fast personalization of text-to-image models

    Rinon Gal, Moab Arar, Yuval Atzmon, Amit H Bermano, Gal Chechik, and Daniel Cohen-Or. Designing an encoder for fast personalization of text-to-image models. arXiv preprint arXiv:2302.12228, 2023. 2

  16. [24]

    Encoder-based domain tuning for fast personalization of text-to-image models

    Rinon Gal, Moab Arar, Yuval Atzmon, Amit H Bermano, Gal Chechik, and Daniel Cohen-Or. Encoder-based domain tuning for fast personalization of text-to-image models. Transactions on Graphics, 42(4):1–13, 2023. 2

  17. [25]

    Talecrafter: Interactive story visualization with multiple characters

    Yuan Gong, Youxin Pang, Xiaodong Cun, Menghan Xia, Haoxin Chen, Longyue Wang, Yong Zhang, Xintao Wang, Ying Shan, and Yujiu Yang. Talecrafter: Interactive story visualization with multiple characters. Siggraph Asia, 2023. 2

  18. [26]

    3d-render-style-lora-sdxl

    GoofyAI. 3d-render-style-lora-sdxl. Hugging Face, https : / / huggingface . co / goofyai / 3d _ render_style_xl. 13

  19. [27]

    Mix-of-show: Decentralized low- rank adaptation for multi-concept customization of diffusion models

    Yuchao Gu, Xintao Wang, Jay Zhangjie Wu, Yujun Shi, Yun- peng Chen, Zihan Fan, Wuyou Xiao, Rui Zhao, Shuning Chang, Weijia Wu, et al. Mix-of-show: Decentralized low- rank adaptation for multi-concept customization of diffusion models. NeurIPS, 2023. 2

  20. [28]

    Densepose: Dense human pose estimation in the wild

    Rıza Alp Güler, Natalia Neverova, and Iasonas Kokkinos. Densepose: Dense human pose estimation in the wild. In CVPR, 2018. 2, 3

  21. [29]

    Svdiff: Compact pa- rameter space for diffusion fine-tuning

    Ligong Han, Yinxiao Li, Han Zhang, Peyman Milanfar, Dimitris Metaxas, and Feng Yang. Svdiff: Compact pa- rameter space for diffusion fine-tuning. arXiv preprint arXiv:2303.11305, 2023. 2

  22. [30]

    Vico: Detail-preserving visual condition for personalized text- to-image generation

    Shaozhe Hao, Kai Han, Shihao Zhao, and Kwan-Yee K Wong. Vico: Detail-preserving visual condition for personalized text- to-image generation. arXiv preprint arXiv:2306.00971, 2023. 2

  23. [31]

    Unipor- trait: A unified framework for identity-preserving single- and multi-human image personalization

    Junjie He, Yifeng Geng, and Liefeng Bo. Unipor- trait: A unified framework for identity-preserving single- and multi-human image personalization. arXiv preprint arXiv:2408.05939, 2024. 2, 3, 5, 7, 13, 18

  24. [32]

    A data perspective on en- hanced identity preservation for diffusion personalization

    Xingzhe He, Zhiwen Cao, Nicholas Kolkin, Lantao Yu, Helge Rhodin, and Ratheesh Kalarot. A data perspective on en- hanced identity preservation for diffusion personalization. arXiv preprint arXiv:2311.04315, 2023. 2

  25. [33]

    Denoising diffu- sion probabilistic models

    Jonathan Ho, Ajay Jain, and Pieter Abbeel. Denoising diffu- sion probabilistic models. NeurIPS, 2020. 2

  26. [34]

    Lora: Low-rank adaptation of large language models

    Edward J Hu, Yelong Shen, Phillip Wallis, Zeyuan Allen-Zhu, Yuanzhi Li, Shean Wang, Lu Wang, and Weizhu Chen. Lora: Low-rank adaptation of large language models. ICLR, 2021. 2, 26

  27. [35]

    Catalin Ionescu, Dragos Papava, Vlad Olaru, and Cristian Sminchisescu. Human3. 6m: Large scale datasets and predic- tive methods for 3d human sensing in natural environments. IEEE transactions on pattern analysis and machine intelli- gence, 36(7):1325–1339, 2013. 7

  28. [36]

    Whole-body human pose estimation in the wild

    Sheng Jin, Lumin Xu, Jin Xu, Can Wang, Wentao Liu, Chen Qian, Wanli Ouyang, and Ping Luo. Whole-body human pose estimation in the wild. In ECCV, 2020. 7

  29. [37]

    Humansd: A native skeleton-guided diffusion model for human image generation

    Xuan Ju, Ailing Zeng, Chenchen Zhao, Jianan Wang, Lei Zhang, and Qiang Xu. Humansd: A native skeleton-guided diffusion model for human image generation. In ICCV, pages 15988–15998, 2023. 2, 3, 8

  30. [38]

    Focusclip: Multimodal subject-level guid- ance for zero-shot transfer in human-centric tasks, 2024

    Muhammad Saif Ullah Khan, Muhammad Ferjad Naeem, Fed- erico Tombari, Luc Van Gool, Didier Stricker, and Muham- mad Zeshan Afzal. Focusclip: Multimodal subject-level guid- ance for zero-shot transfer in human-centric tasks, 2024. 7

  31. [39]

    Datid-3d: Diversity- preserved domain adaptation using text-to-image diffusion for 3d generative model

    Gwanghyun Kim and Se Young Chun. Datid-3d: Diversity- preserved domain adaptation using text-to-image diffusion for 3d generative model. In CVPR, 2023. 2

  32. [40]

    Dif- fusionclip: Text-guided diffusion models for robust image manipulation

    Gwanghyun Kim, Taesung Kwon, and Jong Chul Ye. Dif- fusionclip: Text-guided diffusion models for robust image manipulation. In CVPR, 2022

  33. [41]

    Podia- 3d: Domain adaptation of 3d generative model across large domain gap using pose-preserved text-to-image diffusion

    Gwanghyun Kim, Ji Ha Jang, and Se Young Chun. Podia- 3d: Domain adaptation of 3d generative model across large domain gap using pose-preserved text-to-image diffusion. In CVPR, pages 22603–22612, 2023. 3

  34. [42]

    Beyondscene: Higher-resolution human- centric scene generation with pretrained diffusion

    Gwanghyun Kim, Hayeon Kim, Hoigi Seo, Dong Un Kang, and Se Young Chun. Beyondscene: Higher-resolution human- centric scene generation with pretrained diffusion. In ECCV,

  35. [43]

    Omg: Occlusion-friendly personalized multi-concept generation in diffusion models

    Zhe Kong, Yong Zhang, Tianyu Yang, Tao Wang, Kaihao Zhang, Bizhu Wu, Guanying Chen, Wei Liu, and Wenhan Luo. Omg: Occlusion-friendly personalized multi-concept generation in diffusion models. In ECCV, 2024. 2, 3, 5, 7, 13, 14, 16, 17, 26

  36. [44]

    Inference code for omg + instantid

    Kongzhe. Inference code for omg + instantid. GitHub, https : / / github . com / kongzhecn / OMG / blob / master/inference_instantid.py, 2023. Accessed: 2024-11-20. 26

  37. [45]

    Viescore: Towards explainable metrics for conditional image synthesis evaluation

    Max Ku, Dongfu Jiang, Cong Wei, Xiang Yue, and Wenhu Chen. Viescore: Towards explainable metrics for conditional image synthesis evaluation. arXiv preprint arXiv:2312.14867,

  38. [46]

    Multi-concept customization of text- to-image diffusion

    Nupur Kumari, Bingliang Zhang, Richard Zhang, Eli Shecht- man, and Jun-Yan Zhu. Multi-concept customization of text- to-image diffusion. In CVPR, pages 1931–1941, 2023. 2

  39. [47]

    Applying guidance in a limited interval improves sample and distribution quality in diffusion models

    Tuomas Kynkäänniemi, Miika Aittala, Tero Karras, Samuli Laine, Timo Aila, and Jaakko Lehtinen. Applying guidance in a limited interval improves sample and distribution quality in diffusion models. Advances in Neural Information Processing Systems, 37:122458–122483, 2025. 5

  40. [48]

    Black Forest Labs. Flux. https://github.com/ black-forest-labs/flux, 2024. 8, 13

  41. [49]

    Blip- 2: Bootstrapping language-image pre-training with frozen image encoders and large language models

    Junnan Li, Dongxu Li, Silvio Savarese, and Steven Hoi. Blip- 2: Bootstrapping language-image pre-training with frozen image encoders and large language models. arXiv preprint arXiv:2301.12597, 2023. 7

  42. [50]

    Photomaker: Customizing real- istic human photos via stacked id embedding

    Zhen Li, Mingdeng Cao, Xintao Wang, Zhongang Qi, Ming- Ming Cheng, and Ying Shan. Photomaker: Customizing real- istic human photos via stacked id embedding. arXiv preprint arXiv:2312.04461, 2023. 2, 13, 23

  43. [51]

    Hyperhuman: Hyper-realistic human gener- ation with latent structural diffusion

    Xian Liu, Jian Ren, Aliaksandr Siarohin, Ivan Skorokhodov, Yanyu Li, Dahua Lin, Xihui Liu, Ziwei Liu, and Sergey Tulyakov. Hyperhuman: Hyper-realistic human gener- ation with latent structural diffusion. arXiv preprint arXiv:2310.08579, 2023. 2, 3

  44. [52]

    Cones 2: Customizable image synthesis with multiple subjects

    Zhiheng Liu, Yifei Zhang, Yujun Shen, Kecheng Zheng, Kai Zhu, Ruili Feng, Yu Liu, Deli Zhao, Jingren Zhou, and Yang Cao. Cones 2: Customizable image synthesis with multiple subjects. arXiv preprint arXiv:2305.19327, 2023. 2

  45. [53]

    Matthew Loper, Naureen Mahmood, Javier Romero, Gerard Pons-Moll, and Michael J. Black. SMPL: A skinned multi- person linear model. ACM Transactions Graphics, 34(6): 248:1–248:16, 2015. 3

  46. [54]

    Unified multi-modal latent diffusion for joint subject and text conditional image generation

    Yiyang Ma, Huan Yang, Wenjing Wang, Jianlong Fu, and Jiaying Liu. Unified multi-modal latent diffusion for joint subject and text conditional image generation. arXiv preprint arXiv:2303.09319, 2023. 2

  47. [55]

    Mimo: Controllable character video synthesis with spatial decomposed modeling

    Yifang Men, Yuan Yao, Miaomiao Cui, and Liefeng Bo. Mimo: Controllable character video synthesis with spatial decomposed modeling. arXiv preprint arXiv:2409.16160 ,

  48. [56]

    T2i- adapter: Learning adapters to dig out more controllable ability for text-to-image diffusion models

    Chong Mou, Xintao Wang, Liangbin Xie, Yanze Wu, Jian Zhang, Zhongang Qi, Ying Shan, and Xiaohu Qie. T2i- adapter: Learning adapters to dig out more controllable ability for text-to-image diffusion models. arXiv preprint arXiv:2302.08453, 2023. 8, 13

  49. [57]

    pixel-portraits-lora-sdxl

    Nerijs. pixel-portraits-lora-sdxl. Hugging Face, https:// huggingface.co/nerijs/pixelportraits192- XL-v1.0. 13

  50. [58]

    jojo-style-lora-sdxl

    Norod78. jojo-style-lora-sdxl. Hugging Face, https:// huggingface.co/Norod78/SDXL-JojosoStyle- Lora-v2. 13

  51. [59]

    crayon-style-lora-sdxl

    Ostris. crayon-style-lora-sdxl. Hugging Face, https:// huggingface.co/ostris/crayon_style_lora_ sdxl. 13

  52. [60]

    Cross initialization for personalized text- to-image generation

    Lianyu Pang, Jian Yin, Haoran Xie, Qiping Wang, Qing Li, and Xudong Mao. Cross initialization for personalized text- to-image generation. arXiv preprint arXiv:2312.15905, 2023. 2

  53. [61]

    Georgios Pavlakos, Vasileios Choutas, Nima Ghorbani, Timo Bolkart, Ahmed A. A. Osman, Dimitrios Tzionas, and Michael J. Black. Expressive body capture: 3d hands, face, and body from a single image. In CVPR, 2019. 3, 4, 7, 25

  54. [62]

    Smpl made simple faqs

    MPI for Intelligent Systems Perceiving Systems Department. Smpl made simple faqs. 7

  55. [63]

    Orthogonal adaptation for modular customization of diffusion models

    Ryan Po, Guandao Yang, Kfir Aberman, and Gordon Wet- zstein. Orthogonal adaptation for modular customization of diffusion models. arXiv preprint arXiv:2312.02432, 2023. 3

  56. [64]

    Sdxl: Improving latent diffusion models for high-resolution image synthesis

    Dustin Podell, Zion English, Kyle Lacey, Andreas Blattmann, Tim Dockhorn, Jonas Müller, Joe Penna, and Robin Rombach. Sdxl: Improving latent diffusion models for high-resolution image synthesis. arXiv preprint arXiv:2307.01952, 2023. 7, 8, 13, 26

  57. [65]

    Learning transferable visual models from natural language supervision

    Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, et al. Learning transferable visual models from natural language supervision. In ICML. PMLR, 2021. 2, 7, 25

  58. [66]

    Dreambooth: Fine tuning text-to-image diffusion models for subject-driven gen- eration

    Nataniel Ruiz, Yuanzhen Li, Varun Jampani, Yael Pritch, Michael Rubinstein, and Kfir Aberman. Dreambooth: Fine tuning text-to-image diffusion models for subject-driven gen- eration. In CVPR, 2023. 2, 7, 13, 14, 17, 26

  59. [67]

    Hyperdreambooth: Hypernetworks for fast personalization of text-to-image models

    Nataniel Ruiz, Yuanzhen Li, Varun Jampani, Wei Wei, Tingbo Hou, Yael Pritch, Neal Wadhwa, Michael Rubinstein, and Kfir Aberman. Hyperdreambooth: Hypernetworks for fast personalization of text-to-image models. arXiv preprint arXiv:2307.06949, 2023. 2

  60. [68]

    Eliminating oversaturation and artifacts of high guidance scales in diffusion models

    Seyedmorteza Sadat, Otmar Hilliges, and Romann M We- ber. Eliminating oversaturation and artifacts of high guidance scales in diffusion models. In The Thirteenth International Conference on Learning Representations, 2024. 5

  61. [69]

    Pho- torealistic text-to-image diffusion models with deep language understanding

    Chitwan Saharia, William Chan, Saurabh Saxena, Lala Li, Jay Whang, Emily L Denton, Kamyar Ghasemipour, Raphael Gontijo Lopes, Burcu Karagol Ayan, Tim Salimans, et al. Pho- torealistic text-to-image diffusion models with deep language understanding. NeurIPS, 2022. 2, 5

  62. [70]

    Improved techniques for training gans

    Tim Salimans, Ian Goodfellow, Wojciech Zaremba, Vicki Cheung, Alec Radford, and Xi Chen. Improved techniques for training gans. NeurIPS, 29, 2016. 7

  63. [71]

    Facenet: A unified embedding for face recognition and clus- tering

    Florian Schroff, Dmitry Kalenichenko, and James Philbin. Facenet: A unified embedding for face recognition and clus- tering. In CVPR, pages 815–823, 2015. 7, 25

  64. [72]

    Rethinking the spatial inconsistency in classifier- free diffusion guidance

    Dazhong Shen, Guanglu Song, Zeyue Xue, Fu-Yun Wang, and Yu Liu. Rethinking the spatial inconsistency in classifier- free diffusion guidance. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages 9370–9379, 2024. 5

  65. [73]

    In- stantbooth: Personalized text-to-image generation without test-time finetuning

    Jing Shi, Wei Xiong, Zhe Lin, and Hyun Joon Jung. In- stantbooth: Personalized text-to-image generation without test-time finetuning. arXiv preprint arXiv:2304.03411, 2023. 2

  66. [74]

    Continual diffusion: Continual customization of text-to-image diffusion with c-lora

    James Seale Smith, Yen-Chang Hsu, Lingyu Zhang, Ting Hua, Zsolt Kira, Yilin Shen, and Hongxia Jin. Continual diffusion: Continual customization of text-to-image diffusion with c-lora. arXiv preprint arXiv:2304.06027, 2023. 2

  67. [75]

    Denoising diffusion implicit models

    Jiaming Song, Chenlin Meng, and Stefano Ermon. Denoising diffusion implicit models. In ICLR, 2020. 2 11

  68. [76]

    pokemon-trainer-sprite-pixelart-lora-sdxl

    sWizad. pokemon-trainer-sprite-pixelart-lora-sdxl. Hug- ging Face, https : / / huggingface . co / sWizad / pokemon-trainer-sprite-pixelart . 13

  69. [77]

    Key-locked rank one editing for text-to-image personalization

    Yoad Tewel, Rinon Gal, Gal Chechik, and Yuval Atzmon. Key-locked rank one editing for text-to-image personalization. In ACM SIGGRAPH 2023 Conference Proceedings, pages 1–11, 2023. 2

  70. [78]

    Concept decomposition for visual exploration and inspiration

    Yael Vinker, Andrey V oynov, Daniel Cohen-Or, and Ariel Shamir. Concept decomposition for visual exploration and inspiration. Transactions on Graphics, 2023. 2

  71. [79]

    p+: Extended textual conditioning in text-to-image generation

    Andrey V oynov, Qinghao Chu, Daniel Cohen-Or, and Kfir Aberman. p+: Extended textual conditioning in text-to-image generation. arXiv preprint arXiv:2303.09522, 2023. 2

  72. [80]

    Stable-pose: Lever- aging transformers for pose-guided text-to-image generation

    Jiajun Wang, Morteza Ghahremani Boozandani, Yitong Li, Björn Ommer, and Christian Wachinger. Stable-pose: Lever- aging transformers for pose-guided text-to-image generation. NeurIPS, 37:65670–65698, 2025. 2, 3

  73. [81]

    Instantid: Zero-shot identity-preserving generation in seconds

    Qixun Wang, Xu Bai, Haofan Wang, Zekui Qin, and Anthony Chen. Instantid: Zero-shot identity-preserving generation in seconds. arXiv preprint arXiv:2401.07519, 2024. 2, 4, 5, 7, 13, 16, 17, 23, 25, 26

  74. [82]

    Instantid: Zero-shot identity-preserving generation in seconds, 2024

    Qixun Wang, Xu Bai, Haofan Wang, Zekui Qin, Anthony Chen, Huaxia Li, Xu Tang, and Yao Hu. Instantid: Zero-shot identity-preserving generation in seconds, 2024. 6

  75. [83]

    Ms-diffusion: Multi-subject zero-shot im- age personalization with layout guidance

    X Wang, Siming Fu, Qihan Huang, Wanggui He, and Hao Jiang. Ms-diffusion: Multi-subject zero-shot im- age personalization with layout guidance. arXiv preprint arXiv:2406.07209, 2024. 7, 13, 18

  76. [84]

    Elite: Encoding visual concepts into textual embeddings for customized text-to-image generation

    Yuxiang Wei, Yabo Zhang, Zhilong Ji, Jinfeng Bai, Lei Zhang, and Wangmeng Zuo. Elite: Encoding visual concepts into textual embeddings for customized text-to-image generation. arXiv preprint arXiv:2302.13848, 2023. 2

  77. [85]

    Fastcomposer: Tuning-free multi-subject image generation with localized attention

    Guangxuan Xiao, Tianwei Yin, William T Freeman, Frédo Du- rand, and Song Han. Fastcomposer: Tuning-free multi-subject image generation with localized attention. arXiv preprint arXiv:2305.10431, 2023. 2, 3, 5, 7, 25

  78. [86]

    Fastcomposer: Tuning-free multi- subject image generation with localized attention

    Guangxuan Xiao, Tianwei Yin, William T Freeman, Frédo Durand, and Song Han. Fastcomposer: Tuning-free multi- subject image generation with localized attention. Interna- tional Journal of Computer Vision, pages 1–20, 2024. 7, 13, 18

  79. [87]

    Controlnet-union-sdxl-1.0

    Xinsir. Controlnet-union-sdxl-1.0. Hugging Face, https: / / huggingface . co / xinsir / controlnet - union-sdxl-1.0, 2023. 25

  80. [88]

    Facestudio: Put your face everywhere in seconds.arXiv preprint arXiv:2312.02663,

    Yuxuan Yan, Chi Zhang, Rui Wang, Yichao Zhou, Gege Zhang, Pei Cheng, Gang Yu, and Bin Fu. Facestudio: Put your face everywhere in seconds.arXiv preprint arXiv:2312.02663,

  81. [89]

    Effec- tive whole-body pose estimation with two-stages distillation

    Zhendong Yang, Ailing Zeng, Chun Yuan, and Yu Li. Effec- tive whole-body pose estimation with two-stages distillation. In CVPR, 2023. 3

  82. [90]

    Ip- adapter: Text compatible image prompt adapter for text-to- image diffusion models

    Hu Ye, Jun Zhang, Sibo Liu, Xiao Han, and Wei Yang. Ip- adapter: Text compatible image prompt adapter for text-to- image diffusion models. 2023. 2, 5, 7, 13, 16, 17, 23, 26

  83. [91]

    Adding conditional control to text-to-image diffusion models

    Lvmin Zhang, Anyi Rao, and Maneesh Agrawala. Adding conditional control to text-to-image diffusion models. In CVPR, 2023. 2, 3, 4, 7, 8, 13, 26

  84. [92]

    Composi- tional inversion for stable diffusion models

    Xu-Lu Zhang, Xiao-Yong Wei, Jin-Lin Wu, Tian-Yi Zhang, Zhao-Xiang Zhang, Zhen Lei, and Qing Li. Composi- tional inversion for stable diffusion models. arXiv preprint arXiv:2312.08048, 2023. 2

  85. [93]

    Id-patch: Robust id association for group photo personalization

    Yimeng Zhang, Tiancheng Zhi, Jing Liu, Shen Sang, Liming Jiang, Qing Yan, Sijia Liu, and Linjie Luo. Id-patch: Robust id association for group photo personalization. arXiv preprint arXiv:2411.13632, 2024. 2, 3, 5

  86. [94]

    Catversion: Concatenating embeddings for diffusion-based text-to-image personalization

    Ruoyu Zhao, Mingrui Zhu, Shiyin Dong, Nannan Wang, and Xinbo Gao. Catversion: Concatenating embeddings for diffusion-based text-to-image personalization. arXiv preprint arXiv:2311.14631, 2023. 2

  87. [95]

    Cus- tomization assistant for text-to-image generation

    Yufan Zhou, Ruiyi Zhang, Jiuxiang Gu, and Tong Sun. Cus- tomization assistant for text-to-image generation. arXiv preprint arXiv:2312.03045, 2023. 2

  88. [96]

    En- hancing detail preservation for customized text-to-image generation: A regularization-free approach

    Yufan Zhou, Ruiyi Zhang, Tong Sun, and Jinhui Xu. En- hancing detail preservation for customized text-to-image generation: A regularization-free approach. arXiv preprint arXiv:2305.13579, 2023. 2

  89. [97]

    Storymaker: Towards holistic consistent characters in text-to-image generation

    Zhengguang Zhou, Jing Li, Huaxia Li, Nemo Chen, and Xu Tang. Storymaker: Towards holistic consistent characters in text-to-image generation. arXiv preprint arXiv:2409.12576,

  90. [98]

    overweight,

    Shenhao Zhu, Junming Leo Chen, Zuozhuo Dai, Qingkun Su, Yinghui Xu, Xun Cao, Yao Yao, Hao Zhu, and Siyu Zhu. Champ: Controllable and consistent human image animation with 3d parametric guidance. arXiv preprint arXiv:2403.14781, 2024. 2, 3 12 PersonaCraft: Personalized and Cont...

  91. [99]

    The study collected a total of 18,540 responses from 103 participants across 15 cases, including both custom and COCO-Wholebody scenar- ios

    Naturalness: Rank the images in order of the most natural- looking, considering factors such as physically impossible appearances, illogical features, inconsistencies, or lack of real-world physics and connections. The study collected a total of 18,540 responses from 103 parti...

  92. [100]

    2) Pose Con- sistency: Rank the images based on how well they reflect the given pose input

    Text Correspondence: Rank the images based on how closely they align with the textual description. 2) Pose Con- sistency: Rank the images based on how well they reflect the given pose input. 3) Naturalness: Rank the images in order of the most natural-looking, considering fact...

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.