Pith. sign in

REVIEW 4 major objections 4 minor 20 references

Controllable Clothing: Precise Labels and Generation for Virtual Try-On with Latent Diffusion Models

T0 review · 4 major / 4 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read Adding precise fit labels lets a virtual try-on model generate many versions of one garment instead of a single average fit.

desk verdict Plausible VITON conditioning idea with honest limitations, but the control claim is unvalidated; worth a referee to demand quantitative evidence. read the letter →

arxiv 2608.05834 v1 pith:P4CZXPWT submitted 2026-08-06 cs.CV

Max Rehman Linder This is my paper · ORCID
classification cs.CV
keywords virtualtry-onlatentdiffusionmodelscontrollableimagegenerationIP-adapterT2Iadaptergarmentfitlabelingsegmentation-basedmeasurementmask-conditioned
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This technical report argues that virtual try-on can be made controllable, not just photorealistic. The author's method augments the training data for a latent diffusion try-on model with labels such as sleeve length, shoulder coverage, neckline depth and width, and garment length relative to the hip, then trains small adapters that take these labels as extra conditioning. At inference the user can adjust the labels to change the fit and styling of the generated garment. The report also shows that continuous pixel-space measurements are more useful than one-hot semantic classes, and that an extra-mask boolean can tell the model to regenerate the waist area. If the method holds up, online retailers could show garments with the true fit rather than an idealized or average fit.

What carries the argument

The machinery is the label-injection path around a frozen latent diffusion model. An IP-adapter, which feeds image embeddings into cross-attention, provides the semantic garment embedding, and a trainable MLP/residual module named ControllableClothing adds the concatenated label vector to that embedding. A custom up-block T2I adapter, a light-weight module that does element-wise addition at up-sampling tensors, injects fine detail. The labels themselves come from a pre-processing chain built from open-source segmentation and keypoint models: distances are measured in pixel space from segmentation masks relative to body keypoints such as hips, chest, shoulders, elbow, and wrist, then normalized by body height or shoulder width and standardized across the dataset. Dynamic masks are created in the data loader, and a boolean label tells the model whether an extra mask region below the garment was added.

What would settle it

Hold the garment image, person image, mask, and text prompt fixed, then sweep one continuous label such as shoulder fraction from zero to one or neckline depth from positive to negative across several random seeds; if the generated images show no systematic change in shoulder coverage or neckline, the label channel is not learned. A second check is to retrain with label values randomly shuffled and compare outputs to the trained model; indistinguishable outputs would mean the label vector is being ignored.

Watch

Extended reading notes

Core claim

The central claim is that pairing garment images with precise fit labels lets a pretrained latent diffusion model generate a diverse, user-steerable set of try-on images, instead of a single average fit. The report constructs a label vector from one-hot classes and continuous floats computed in pixel space—shoulder fraction, hip distance, normalized neckline depth and width, and sleeve-length class—and injects it into an IP-adapter embedding through a small trainable residual module, while a custom up-block T2I adapter supplies detail. Qualitative experiments show that changing the continuous values changes shoulder styling, neckline depth, garment length, and waist regeneration, and that the numeric labels behave better than the one-hot ones. The author is careful to call the evidence preliminary: the up-block adapter was trained for only a fraction of the total epochs, and the unique contribution of the one-hot encoding is unclear because garment type and sleeve length were already supplied as text.

Load-bearing premise

The load-bearing premise is that the custom label vector has its own influence on what the diffusion model generates; if the model already gets the same information from the text prompt and the garment image, the whole control mechanism adds nothing.

Editorial extensions

If this is right

  • If the method works as described, a virtual try-on system can expose controls for shoulder coverage, neckline depth, garment length relative to the hip, and sleeve length, and each control changes the generated outfit rather than only the prompt.
  • Because the position labels are measured in pixel space and normalized, they give the model explicit geometric guidance that a latent text or image embedding alone does not carry; the paper states this is why these features work well despite little training.
  • The mask-plus-boolean mechanism makes it possible to regenerate a waist or crop-top area upward from the mask boundary, letting a user dress a person in a shorter garment than the one in the source photo.
  • The paper's critique of one-hot semantic labels implies that future versions should encode attributes such as sleeve length as continuous variables between zero and one.

Reading between the lines

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

  • A natural ablation the paper did not report would train with the label vector present versus the same labels removed or shuffled while keeping text prompts fixed; near-identical outputs would mean the claimed control mechanism reduces to the text prompt.
  • Because the current labels are global scalars, the architecture cannot pin a boundary to an exact pixel row; combining label injection with spatial conditioning such as a hemline mask is a likely next step for precise geometric placement.
  • The segmentation-based measurement pipeline is itself a load-bearing contribution, so the method's reliability is tied to segmentation quality; an extension would test robustness by adding controlled noise to shoulder and neckline masks during training.
  • If the waist-mask boolean generalizes, the same boundary-regeneration trick could apply to necklines, cuffs, or hemlines, turning the model into a general 'regenerate from this boundary' tool for garment edits.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 4 minor

Summary. The paper proposes ControllableClothing, a method for adding fine-grained control to latent-diffusion-based virtual try-on. Garment images are pre-processed with segmentation (GroundingDINO, SAM, OpenPose) to extract continuous measurements (shoulder coverage fraction, hip distance, neckline depth/width) and one-hot class labels (sleeve length, fit, style). These labels are concatenated with an IP-adapter image embedding and passed through a custom trainable MLP; additionally, a custom up-block T2I adapter is introduced. Training on DressCode and VitonHD, the author claims the model can steer fit and style at inference. Evaluation is qualitative: a handful of generated images showing changed conditions, plus a one-image ablation. The author openly acknowledges in Section 8 that the one-hot encoding effect is unclear and that several labels were also passed as text prompts during training.

Significance. If the central claim holds, the method would address a real limitation of VITON systems, which typically produce a single average fit with no user control over sleeve length, neckline depth, or shoulder coverage. The proposed label-extraction pipeline in pixel space is a plausible and reproducible alternative to latent-space captioning, and the custom up-block T2I adapter is a sensible architectural contribution. The author's decision to train on open datasets with open tools and to document limitations honestly (Section 8) counts in the paper's favor. However, the reported evidence does not currently establish the causal control claim, because the evaluation lacks quantitative metrics, seed control, repeated generations, and a held-out split. The paper is therefore best seen as an early technical report with a promising direction rather than a validated method.

major comments (4)
  1. [Section 6, Figures 14–20] The central claim that user-controllable labels steer generation is supported only by a small set of curated single images. No inference seeds are reported, no repeated generations per condition are shown, and no quantitative metric (e.g., LPIPS, FID, measured garment-length error, or fraction of correct semantic labels) distinguishes label-driven variation from sampling noise. Because Figures 14–15 are generated without text prompts, the differences could reflect prompt-free behavior or seed variation rather than the one-hot vector. I request a controlled evaluation: fix the seed, mask, prompt, and all other inputs, vary one label at a time, and report distributions of a relevant metric over multiple seeds, ideally on a held-out split of DressCode/VitonHD rather than on the training set.
  2. [Section 8, Discussion of one-hot encoding] The manuscript itself states, 'the effect of the one-hot encoding is unclear' and that 'the type of garment ... was passed as a prompt along with five other random labels' and that sleeve length information was 'added as a prompt.' Since the text prompt already conveyed these attributes, the label vector's non-redundant influence is exactly the load-bearing mechanism asserted in the abstract. The ablation study in Section 7 removes entire modules but does not test whether the label input alone changes the output when the prompt is held fixed. To support the control claim, the author needs an experiment where the label vector is varied while the text prompt (if any) and all other conditioning are held constant; without such a test, the method may reduce to prompt conditioning plus sampling noise.
  3. [Section 6.3, Figure 19] The shoulder-fraction demonstration shows a single pair of images with left-frac and right-frac set to different values. This is a suggestive example, but a single unseeded pair cannot exclude cherry-picking, especially because the paper reports that manual changes can have 'unforeseen consequences' (Section 6.1, Figure 14). For a claim that 'setting right-frac to 1 and left to 0' yields asymmetric shoulder coverage, I ask for several seeds per condition and a quantitative measure of shoulder coverage in the output, computed with the same segmentation pipeline used for labels.
  4. [Section 6.4, Figure 20] The waist-mask control experiment also relies on two examples per condition without seeds or a numeric measure of the resulting garment length. Since the comparison is between two generated images with 'the exact same mask' but different waist-mask values, the relevant question is whether this parameter changes the output distribution systematically. Please provide repeated generations and a metric such as the distance from the hem to the mask boundary.
minor comments (4)
  1. [Section 6.1] The text references 'Figure 24' in the discussion of Figure 13, and the figure numbering in the appendix appears inconsistent (e.g., 'Figure 24' is described as a 'relatively good example' while the appendix orders figures differently); please renumber and cross-check all figure references.
  2. [Section 4.4] Table 1 lists questions and answers from GPT-4o, but the reader cannot tell which of these were actually used as labels versus which were rejected; Section 8 mentions that some question types were rejected. Please clarify the final label set and its encoding in a single table.
  3. [Section 5.2] The dimensions of the up-block T2I intermediate tensors are listed, but the paper does not give the number of training steps, learning rate, batch size, or the exact random chance for adding extra waist mask. Adding these hyperparameters would improve reproducibility.
  4. [Throughout] There are several typos and grammatical issues (e.g., 'bu' for 'but', 'the model with not learn', 'looses' for 'loses', 'unpainted' for 'unmasked', 'nlp' for 'MLP' in Section 8). A careful proofread is needed.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the control claim is an unproven empirical hypothesis; the report's own Section 8 admits the one-hot conditioning may be redundant, but nothing in the derivation collapses into its inputs.

full rationale

The paper contains no derivation chain whose output equals its input by construction. The diffusion loss (Eq. 3) and the label-conditioned adapters are standard conditional modeling; the trained MLP is not constrained so that varying a label at inference reproduces the label's training-time measurement. The continuous labels (shoulder fraction, hip distance, neckline depth) are extracted by pixel-space segmentation and are then used as conditioning inputs; there is no fitted parameter that is later renamed as a 'prediction' of the same quantity. Likewise, the central claim that adapters enable user control rests on qualitative images, not on an equation that forces the result. The most serious admitted limitation appears in Section 8: 'the effect of the one-hot encoding is unclear' and garment type and sleeve length 'was passed as a prompt along with five other random labels.' This is a genuine threat to the causal interpretation of the control claim—the one-hot vector may be redundant with text conditioning—but redundancy is a confound, not a circular reduction. The report also evaluates on images from the training distribution without a held-out split or seed controls, which undermines the evidence for the control claim as an empirical matter. However, circularity requires exhibiting a specific reduction of a claimed result to its own inputs or to a load-bearing self-citation; no such reduction is present, and no load-bearing result is imported from the author's own prior work. The score is therefore 0.

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

The central claim rests on the adequacy of the label representation, the accuracy of the measurement stack, and the learnability of the adapters. These are domain assumptions with only qualitative support. No new physical entities are introduced.

assumptions (5)
  • ad hoc to paper The chosen scalar and class labels can summarize the aspects of garment fit and style that matter for control.
    The paper selects shoulder fraction, neckline depth and width, hip distance, and sleeve-length classes without evidence that these suffice for user-relevant control; this is the premise of the whole method.
  • domain assumption The segmentation and keypoint models yield measurements accurate enough for training labels.
    All fit labels derive from GroundingDINO, SAM, and OpenPose outputs (Section 4.3.2); no error analysis is provided.
  • ad hoc to paper The custom MLP and T2I adapters can learn to map the label vector to output changes.
    This is the learnability assumption behind ControllableClothing; the report only shows a few qualitative examples and admits the up-block T2I was trained for only 7 epochs.
  • domain assumption SDXL and the pretrained IP-adapter provide a suitable base for VITON fine-tuning.
    The method reuses these open-source components and does not validate their suitability beyond the reported examples.
  • standard math The diffusion process obeys the standard DDPM equations (1)-(5).
    Standard formulation from Ho et al. (2020); this is background, not in dispute.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Controllable Clothing: Precise Labels and Generation for Virtual Try-On with Latent Diffusion Models." pith.science (2026). https://pith.science/paper/P4CZXPWT

@misc{pith2026260805834,
  author       = {Pith},
  title        = {Pith review of: Controllable Clothing: Precise Labels and Generation for Virtual Try-On with Latent Diffusion Models},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/P4CZXPWT}},
  note         = {Machine review of arXiv:2608.05834}
}
read the original abstract

In this technical report, I present a new method for guiding image generation in the context of Virtual- Try-On (VITON). The proposed method leverages new open source Ai models to augment the image data with labels, such as lengths and styles. By training adapters with these labels paired with images of the garments, the model can produce a more diverse set of images that the user can control. For the end user, such as a retailer, this means that they can assure that the produced image is as true to the true fit as possible, not misleading consumers

Figures

Figures reproduced from arXiv: 2608.05834 by the authors.

Figure 1
Figure 1. How in-painting can be used to create photo realistic images 3.2 Diffusion models for Virtual Try-On 3.2.1 TryOnDiffusion The first published paper on diffusion models for virtual try-on published summer of 2023 (to my knowl￾edge), is TryOnDiffusion by Google [13] [PITH_FULL_IMAGE:figures/full_fig_p005_1.png] view at source ↗
Figure 2
Figure 2. The architecture of TryOnDiffusion uses parallel unets in the diffusion process. This model were the first that showed reliably "good" results, and compared to all other models that it was compared to (all GANs at the time), it greatly outperformed everyone of them. This model was never released publicly and the massive dataset of 4 millions images were scraped of the internet. So whereas this model could never be t… view at source ↗
Figure 3
Figure 3. The architecture of IDM. Similarly to OOTD, it uses parallel unets and concatenates self-attention maps before doing the self-attention. The main difference between the two is firstly the prevalence of an ip-adapter at IDM viton, but also the fact that the non-denoising unet is frozen. Having a frozen pre-trained image-to-image unet appears to be a good direction. Intuitively, the frozen unet maintains details about… view at source ↗
Figures from the paper (22 more)
Figure 4
Figure 4. Figure 4: Example inference of IDM-viton, where the garment is mapped onto the masked area of the person. Note that it preserves details but also generates areas on arms such as the hem on arms and the forearms in their entirety. In short, almost 2 years after stable diffusion’s…
Figure 5
Figure 5. Figure 5: The CLIP text embedding that SDXL is trained with a corresponding CLIP image-embedding Since the text and image prompt are contrasted to each other (maximizes dot-product), the embedding corresponding to an prompt of a description of a garment will by default be very s…
Figure 6
Figure 6. Figure 6: Visualisation of garment length The amount of bare shoulders was decided by measuring if there was segmented garment on the shoul￾ders. At the key-point, a normalized radius was taken and the fraction of pixels that was on the garment was measured. fraction = garment a…
Figure 7
Figure 7. Figure 7: The green and red lines measure depth and widht of neckline Length of sleeves was not measured as a distance, but rather as one-hot encoded classes with 6 varying lengths. It was done by: • Identify arm with most skin showed by area, some arms may be hidden from view a…
Figure 8
Figure 8. Figure 8: Red indicates that garment was detected It is possible that it would be possible that using the previous techniques, i.e continuous measurement would have worked. But having two different method of measuring length allows for more evaluation. A big advantage is with co…
Figure 9
Figure 9. Figure 9: Example of masked Image What the masking operation also does is to allow for adding extra mask below the waist. If the mask ends precisely at garment model learns to follow line of mask, which may or may not be desirable. For this reason the model is also given informa…
Figure 10
Figure 10. Figure 10: Reference image in dataset Question Answer How is the upper garment fitted around arms? Relaxed How is the upper garment fitted around chest/shoulders for its type? Relaxed How is the upper garment fitted around waistline? Relaxed Where does sleeve-hem or cuffs extend…
Figure 11
Figure 11. Figure 11: The architecture of controllable clothing. The trainable components are the T2I adapters and the fine-tuning MLP for the IP-adapter, that also incorporates the labels and information about the mask. 12 [PITH_FULL_IMAGE:figures/full_fig_p013_11.png]
Figure 12
Figure 12. Figure 12: MSE-Loss over Epochs, note that the VITON dataset was added at epoch 17 which lead to the big increase in loss 13 [PITH_FULL_IMAGE:figures/full_fig_p014_12.png]
Figure 13
Figure 13. Figure 13: Example VITON using controllable clothing [PITH_FULL_IMAGE:figures/full_fig_p015_13.png]
Figure 14
Figure 14. Figure 14: Garment Image Condition Value Sleeve Length Elbow-length Left Frac 0.0 Right Frac 0.0 Left Hip Dist 0.0 Right Hip Dist 0.0 Normalized Neckline Depth 1 Normalized Neckline Width 1 Arm Fit Sleeves are relaxed fit around the arms Waist Fit Relaxed fit around the waist Bu…
Figure 15
Figure 15. Figure 15: Garment Image Condition Value Sleeve Length Full-length extended sleeves Left Frac 0.0 Right Frac 0.0 Left Hip Dist 10 Right Hip Dist 10 Normalized Neckline Depth -2 Normalized Neckline Width -2 Arm Fit Sleeves are tight fit around the arms Chest/Shoulder Fit Tight fi…
Figure 16
Figure 16. Figure 16: Same labels as [PITH_FULL_IMAGE:figures/full_fig_p016_16.png]
Figure 17
Figure 17. Figure 17: For highly detailed garments, detail may be lost unless there is a more trained module for up-blocks in place. interestingly, the image on the right have colours wrong but the right motive [PITH_FULL_IMAGE:figures/full_fig_p017_17.png]
Figure 18
Figure 18. Figure 18: On the other hand, for more plain images, without much intricate detail, the model may produce images. The model need more training to get to a state where it reproduces detail. 6.3 Shoulder fraction [PITH_FULL_IMAGE:figures/full_fig_p017_18.png]
Figure 19
Figure 19. Figure 19: Setting different fractions for shoulders yields different styling of off-shoulder garments Recall that the two parameters frac, left an right, corresponds to percentage of shoulder that is covered in garment. Observing [PITH_FULL_IMAGE:figures/full_fig_p017_19.png]
Figure 20
Figure 20. Figure 20: Setting the parameter waist-mask to true or false yields different generation around waist. Another parameter for inference is waist-mask. In training, with a certain chance, extra mask is added to the image to the waist, forcing the model to regenerate the waist area…
Figure 21
Figure 21. Figure 21: From left to right, we have Original garment, generated image using all components, no up-block T2I, no IP-adapter and no down-block T2I 17 [PITH_FULL_IMAGE:figures/full_fig_p018_21.png]
Figure 22
Figure 22. Figure 22: After a few epochs it became apparent that there was no more learning to be done as the model regenerated artefacts that should be hidden. Like this scarf on the left, that is completely covered by the mask. 19 [PITH_FULL_IMAGE:figures/full_fig_p020_22.png]
Figure 23
Figure 23. Figure 23: Different styles of outfitting a garment 21 [PITH_FULL_IMAGE:figures/full_fig_p022_23.png]
Figure 24
Figure 24. Figure 24: Different styles of outfitting a garment [PITH_FULL_IMAGE:figures/full_fig_p023_24.png]
Figure 25
Figure 25. Figure 25: Further example of how changing shoulder-frac may change type of garment 22 [PITH_FULL_IMAGE:figures/full_fig_p023_25.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

20 extracted references · 14 canonical work pages

  1. [1]

    High- resolution image synthesis with latent diffusion models, 2022

    Robin Rombach, Andreas Blattmann, Dominik Lorenz, Patrick Esser, and Björn Ommer. High- resolution image synthesis with latent diffusion models, 2022

    work page 2022

  2. [2]

    Sdxl: Improving latent diffusion models for high-resolution image synthesis, 2023

    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, 2023

    2023

  3. [3]

    Imagdressing-v1: Customizable virtual dressing, 2024

    Fei Shen, Xin Jiang, Xin He, Hu Ye, Cong Wang, Xiaoyu Du, Zechao Li, and Jinhui Tang. Imagdressing-v1: Customizable virtual dressing, 2024

    work page 2024

  4. [4]

    Viton-gan: Virtual try-on image generator trained with adversarial loss.Eurograph- ics 2019 - Posters, 2019

    Shion Honda. Viton-gan: Virtual try-on image generator trained with adversarial loss.Eurograph- ics 2019 - Posters, 2019

    work page 2019

  5. [5]

    Viton-hd: High-resolution virtual try-on via misalignment-aware normalization, 2021

    Seunghwan Choi, Sunghyun Park, Minsoo Lee, and Jaegul Choo. Viton-hd: High-resolution virtual try-on via misalignment-aware normalization, 2021

    work page 2021

  6. [7]

    Adding conditional control to text-to-image diffusion models, 2023

    Lvmin Zhang, Anyi Rao, and Maneesh Agrawala. Adding conditional control to text-to-image diffusion models, 2023

    2023

  7. [8]

    Animate anyone: Consistent and controllable image-to-video synthesis for character animation, 2024

    Li Hu, Xin Gao, Peng Zhang, Ke Sun, Bang Zhang, and Liefeng Bo. Animate anyone: Consistent and controllable image-to-video synthesis for character animation, 2024

    work page 2024

  8. [9]

    Improving diffusion models for authentic virtual try-on in the wild, 2024

    Yisol Choi, Sangkyung Kwak, Kyungmin Lee, Hyungwon Choi, and Jinwoo Shin. Improving diffusion models for authentic virtual try-on in the wild, 2024

    work page 2024

Show all 20 references
  1. [10]

    Stableviton: Learning semantic correspondence with latent diffusion model for virtual try-on, 2023

    Jeongho Kim, Gyojung Gu, Minho Park, Sunghyun Park, and Jaegul Choo. Stableviton: Learning semantic correspondence with latent diffusion model for virtual try-on, 2023

    2023

  2. [11]

    Tryondiffusion: A tale of two unets, 2023

    Luyang Zhu, Dawei Yang, Tyler Zhu, Fitsum Reda, William Chan, Chitwan Saharia, Mohammad Norouzi, and Ira Kemelmacher-Shlizerman. Tryondiffusion: A tale of two unets, 2023

    2023

  3. [12]

    Ip-adapter: Text compatible image prompt adapter for text-to-image diffusion models, 2023

    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

    2023

  4. [13]

    Learning transferable visual models from natural language supervision, 2021

    Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, Gretchen Krueger, and Ilya Sutskever. Learning transferable visual models from natural language supervision, 2021

    2021

  5. [14]

    Ootdiffusion: Outfitting fusion based latent diffusion for controllable virtual try-on, 2024

    Yuhao Xu, Tao Gu, Weifeng Chen, and Chengcai Chen. Ootdiffusion: Outfitting fusion based latent diffusion for controllable virtual try-on, 2024

    2024

  6. [15]

    Denoising diffusion probabilistic models, 2020

    Jonathan Ho, Ajay Jain, and Pieter Abbeel. Denoising diffusion probabilistic models, 2020

    2020

  7. [16]

    T2i-adapter: Learning adapters to dig out more controllable ability for text-to-image diffusion models, 2023

    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, 2023

    2023

  8. [17]

    Self-correction for human parsing.arXiv preprint arXiv:1910.09777, 2019

    Peike Li, Yunqiu Xu, Yunchao Wei, and Yi Yang. Self-correction for human parsing.arXiv preprint arXiv:1910.09777, 2019

    1910 arXiv

  9. [18]

    Dress code: High-resolution multi-category virtual try-on, 2022

    Davide Morelli, Matteo Fincato, Marcella Cornia, Federico Landi, Fabio Cesari, and Rita Cuc- chiara. Dress code: High-resolution multi-category virtual try-on, 2022

    2022

  10. [19]

    Grounding dino: Marrying dino with grounded pre-training for open-set object detection, 2024

    Shilong Liu, Zhaoyang Zeng, Tianhe Ren, Feng Li, Hao Zhang, Jie Yang, Qing Jiang, Chunyuan Li, Jianwei Yang, Hang Su, Jun Zhu, and Lei Zhang. Grounding dino: Marrying dino with grounded pre-training for open-set object detection, 2024. 20

    2024

  11. [20]

    Berg, Wan-Yen Lo, Piotr Dollár, and Ross Girshick

    Alexander Kirillov, Eric Mintun, Nikhila Ravi, Hanzi Mao, Chloe Rolland, Laura Gustafson, Tete Xiao, Spencer Whitehead, Alexander C. Berg, Wan-Yen Lo, Piotr Dollár, and Ross Girshick. Seg- ment anything, 2023

    2023

  12. [21]

    G. M. Merz, Y . Liu, C. J. Burke, P. D. Aleo, X. Liu, M. C. Kind, V . Kindratenko, and Y . Liu. Detection, instance segmentation, and classification for astronomical surveys with deep learning (deepdisc): Detectron2 implementation and demonstration with hyper suprime-cam data,...

    2023

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.