Pith. sign in

REVIEW 5 major objections 6 minor 22 references

Uncertainty-Aware Domain Adaptation for Vitiligo Segmentation in Clinical Photographs

T0 review · 5 major / 6 minor · reviewed 2026-08-03 · deepseek-v4-flash

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

Pith's one-line read A frequency-gated, domain-adapted ensemble segments vitiligo in clinical photos with 85.05% Dice and zero missed lesions.

desk verdict Credible vitiligo segmentation paper with useful uncertainty maps, but the core loss is underspecified (no source for the skin mask) and the headline boundary improvement lacks statistical support. read the letter →

arxiv 2512.11791 v2 pith:BI5FFSMH submitted 2025-12-12 cs.CV

Wentao Jiang , Vamsi Varra , Caitlin Perez-Stable , Harrison Zhu , Meredith Apicella , Nicole Nyamongo This is my paper · ORCID
classification cs.CV MSC 68T4568U1092C55 PACS 87.57.nj
keywords vitiligosegmentationclinicalphotographyfrequency-domainanalysisdomainadaptationuncertaintyquantificationskinlesionboundaryerrorensembleinference
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This paper claims that vitiligo lesions in ordinary clinical photographs can be segmented more accurately and more safely than prior baselines by combining three ideas: a Fourier-domain module that amplifies high-frequency boundary textures, a training loss restricted to skin regions so background pixels cannot dominate gradients, and an inference pipeline that averages predictions across multiple model folds and image transformations to produce pixel-level uncertainty maps. On a held-out set of 118 patient-disjoint clinical images, the full framework reaches a mean Dice score of 85.05%, reduces the 95% Hausdorff boundary distance from about 45 px to 30 px, and never completely misses a lesion (0.0% failure rate). The authors argue the clinical value is not just the accuracy gain but the availability of interpretable entropy maps that tell a clinician where to look for unreliable predictions. A sympathetic reader would care because vitiligo extent drives treatment decisions, and current manual assessment is subjective and inconsistent.

What carries the argument

The two load-bearing design elements are the High-Frequency Spectral Gating (HFSG) module and the ROI-constrained dual-task loss. HFSG takes a spatial feature map, applies a 2D real fast Fourier transform, multiplies the spectrum by a static high-pass mask and a learned per-channel gate, transforms back with an inverse FFT, and fuses the result with the original features via channel attention and a residual connection; this is what the paper claims recovers boundary harmonics that low-pass convolutional operators lose. The ROI loss uses a binary skin mask M_skin to zero out losses on non-skin pixels (masked focal and masked Dice) and adds a background-suppression cross-entropy term, so gradi

What would settle it

Run the described pipeline on the same development/test split with and without the skin-mask term (replacing M_skin with an all-ones mask) and with and without the learnable gate in HFSG. If Dice and HD95 on the 118-image test set move by less than the paper's ablation margins, the core mechanisms are not the cause of the reported performance. Additionally, check the public clinical dataset for any skin-label or pretrained-segmenter dependency: the absence of either would make the method irreproducible as written.

Watch

Extended reading notes

Core claim

On its own terms, the central claim is that a segmentation network built on a modern convolutional encoder, initialized on a large public dermoscopy dataset, and augmented with a High-Frequency Spectral Gating module can resolve the fuzzy, depigmented borders of vitiligo lesions in 'in-the-wild' clinical photographs better than standard CNNs and Transformer encoders. The paper reports Dice 85.05% versus 84.07% for a heavyweight Transformer baseline, HD95 of 29.95 px versus 30.90 px, and zero catastrophic failures on their 118-image test set, with ablation results attributing most of the gain to the dermoscopy pre-training (11.5 Dice points) and a smaller but consistent gain to the spectral g

Load-bearing premise

The training objective in Eqs. (4)–(5) requires a binary skin mask M_skin at every training pixel, and the paper never states where those masks come from or how they are supervised; if no reproducible skin-mask source exists, the ROI-constrained loss cannot be implemented as written and the reported gains from 'anatomy-guided' training are unverifiable.

Editorial extensions

If this is right

  • If correct, automated vitiligo extent scoring could become objective and repeatable, replacing subjective VASI-style visual estimation in treatment monitoring.
  • Clinicians could adopt the entropy and variance maps as a human-in-the-loop review screen, focusing attention only on high-uncertainty boundary regions rather than re-checking every pixel.
  • The dermoscopy-to-photography transfer recipe, if reproducible, suggests the same strategy may help other low-annotation skin conditions with fuzzy boundaries.
  • The zero-failure result, if it holds on larger cohorts, would make automated screening of vitiligo photographs safer than current baselines that occasionally miss whole lesions.

Reading between the lines

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

  • The paper leaves open a testable extension: use the predicted entropy maps as an active-learning signal to label only high-uncertainty regions, which could cut annotation cost while improving boundary accuracy; this is my inference, not stated in the paper.
  • If the skin mask M_skin in Eqs. (4)–(5) is in practice derived from the vitiligo ground truth (e.g., skin as the complement of the lesion), the 'anatomy-guided' claim would reduce to a standard masked loss and the mechanism would deserve re-examination; this is an inference based on the paper's silence.
  • A direct cross-check of the central mechanism would be to compare Dice and HD95 with and without the static high-pass mask in HFSG: if gating all frequencies gives the same result, the 'high-frequency' explanation is not the active ingredient.
  • The entropy maps could also be repurposed for unsupervised test-time adaptation, which the authors mention as future work; an immediate test is whether entropy-guided pseudo-labels improve zero-shot performance on dermoscopic images.
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

5 major / 6 minor

Summary. The paper proposes a three-pillar framework for vitiligo segmentation in clinical photographs: (1) domain-adaptive pretraining on ISIC 2019 plus an ROI-constrained dual-task loss, (2) a ConvNeXt V2 encoder augmented with a High-Frequency Spectral Gating (HFSG) module and stem-skip connections, and (3) K-fold ensemble with Test-Time Augmentation to produce pixel-wise uncertainty maps. On an expert-annotated clinical cohort with a patient-level split, the full framework reports a mean Dice of 85.05%, HD95 of 29.95 px, and a 0.0% failure rate on a 118-image test set, outperforming CNN and Transformer baselines. Ablations on a single fold and a qualitative zero-shot dermoscopy experiment are used to attribute gains to the proposed components.

Significance. If the results are reproducible, the paper addresses a relevant clinical problem with a practical deployment angle: automated vitiligo area measurement from in-the-wild photographs, where boundary fuzziness and background clutter are real obstacles. The use of a patient-level split, public ISIC data, K-fold ensembling, and TTA-based uncertainty maps are sensible design choices, and the reported uncertainty visualization is a useful addition for human-in-the-loop review. However, the central training loss depends on a skin mask that is never specified, and the headline statistical claims are not supported by significance testing or confidence intervals. The contribution is promising but needs substantial revision before the performance claims can be taken at face value.

major comments (5)
  1. [§2.3, Eq. (4)–(5)] The ROI-constrained dual-task loss uses a binary skin mask M_skin and an auxiliary skin loss L_skin_aux, but the manuscript never specifies how M_skin is obtained, supervised, or generated. The dataset description in §3.1 only mentions vitiligo lesion annotations; no skin segmenter, thresholding rule, or anatomical prior is cited or described. Without a definition of M_skin, Eq. (5) cannot be computed, so the entire training strategy is not reproducible as written. This is load-bearing, because the claimed 'anatomy-guided hard negative mining' depends on M_skin being a valid skin mask rather than something derived from the vitiligo ground truth (which would leak location information) or an undisclosed external model. Please specify the source of M_skin for every training and test image, including whether it is available during ISIC pretraining, and provide the exact formula for L_skin_au
  2. [§3.3, Table 1] The text states a 'significant reduction in boundary error,' but no confidence intervals, paired significance tests, or effect-size measures are reported. The absolute HD95 improvement over the MiT-B5 baseline is 30.90 px to 29.95 px (0.95 px), which is far smaller than the reported standard deviations (28.83–31.79 px). The claim that the method 'consistently outperforms' all baselines is also not supported in every metric: the proposed method has slightly lower recall (86.70) than MiT-B2 (86.77) and MiT-B5 (86.76). The 0.0% failure rate is likewise asserted without a statistical basis; given that the baseline failure rate is 0.8% on 118 images, this corresponds to roughly one image, and the difference is not testable with the presented evidence. Please report per-image paired comparisons, confidence intervals, and a clear failure definition.
  3. [§3.4, Table 2] The ablation study is conducted on Fold 0 only, and the evidence for the HFSG module's contribution is mixed. Comparing M3 (ISIC + ASPP) with M5 (ISIC + ASPP + HFSG), HD95 worsens from 29.46 px to 30.76 px, while Dice improves only from 84.56 to 84.72. The claim that HFSG 'effectively captures weak spectral signals' and reduces boundary error is therefore not supported by the table's own HD95 numbers. The only apparent improvement in M5 over M3 is the failure rate (0.8% to 0.0%), which is a single-image difference and not statistically meaningful without a defined failure criterion and more cases. Please provide a multi-fold ablation or per-image paired analysis and avoid attributing boundary-error improvements to HFSG based on this table.
  4. [§3.1, §3.2, Table 1] The relationship between the 5-fold cross-validation on the development set and the reported test-set results is unclear. Table 1 says 'Results are reported as Mean±Std over 5-fold cross-validation' on the 'Clinical Test Set,' but the test set is a held-out 15% of patients. Please clarify: were the five fold models also evaluated on the held-out test set, and are the reported numbers the mean and standard deviation over those five models? Also state how the ensemble prediction in Eq. (6) relates to the numbers in Table 1 — is the ensemble evaluated as a single system, or are the five models averaged into the reported mean? This distinction is essential for interpreting the performance claims and for reproducibility.
  5. [§2.2.1, §3.2] The domain-adaptive pretraining on ISIC 2019 is a key component, but the pretraining procedure is not specified. The text only says the ConvNeXt V2 encoder is 'explicitly pre-trained' on ISIC 2019; it does not state the auxiliary task (e.g., classification or segmentation), the loss, the pretraining resolution, the number of epochs, or how the nine-class ISIC labels are used. This makes the 'domain-adaptive pre-training' pillar impossible to reproduce. Please provide the pretraining protocol or cite a specific checkpoint and fine-tuning procedure.
minor comments (6)
  1. [Figure 2] The figure caption and diagram use 'HFSC module' in one label and 'HFSG module' in the text; unify the acronym. Also, the 'optional' low-frequency channel gate is not described in the text.
  2. [§2.3.1] L_bg is introduced in the text but never defined with an equation. Please give the explicit form of the background-suppression term and its weighting in Eq. (5).
  3. [§2.2.2, Eq. (2)] The static high-pass mask M_high is not defined; the cutoff frequency and whether it is applied per-channel or globally are unspecified. The dimension and initialization of W_gate also need clarification.
  4. [§3.1] The 198 'negative patches' in the development set are mentioned but not described: are they image crops, synthetic images, or additional training samples? How are they used in the loss? This matters for data handling and for interpreting the 'artificial negatives were excluded' statement for the test set.
  5. [§3.4, Table 2] The failure rate is reported as 0.8% and 0.0% but no definition of 'failure' is given (e.g., zero Dice, Dice below a threshold, or empty prediction). Define the criterion and report the raw count.
  6. [§3.5] The external dermoscopy evaluation is qualitative only, which the paper acknowledges. It would strengthen the paper to report quantitative results if any annotations exist, or to clearly label this as a motivating case study rather than a validation.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the performance claims rest on held-out test evaluation and external pre-training, with no load-bearing self-citation chain or fit-as-prediction step.

full rationale

I walked the paper's derivation chain from architecture (Eq. 1-3), through the dual-task loss (Eq. 4-5), to inference/uncertainty (Eq. 6-7) and the reported test-set metrics (Table 1). The HFSG module uses a static high-pass mask plus a learnable channel gate; the gate is trained on the development folds, not fitted to the test set, so the reported Dice/HD95 are not constructed from the target labels. The domain-adaptive pre-training on ISIC 2019 is an external, publicly available source and is not derived from the clinical test cohort. The TTA/K-fold ensemble is a standard aggregation over fixed transformations and model folds; the entropy maps are descriptive and are not claimed to be derived from the ground truth. The paper contains no self-citation chain used to justify the central claim; all cited prior work is external. The main substantive gap is that the binary skin mask M_skin and the auxiliary skin loss L_skin_aux in Eq. (4)-(5) are never specified: the paper does not say how M_skin is obtained or supervised. That is a reproducibility/omission concern rather than a demonstrated circular reduction; nothing in the text shows that M_skin is a function of the vitiligo ground truth, nor that the reported test metrics are equivalent by construction to the loss inputs. The paper also honestly states a limitation for the external dermoscopic cohort: 'the absence of pixel-level annotations for the external dermoscopic cohort, restricting us to qualitative validation.' That is a stated limitation, not hidden circularity. Since no prediction is shown to reduce to a fitted parameter or to an author-imported uniqueness claim, the appropriate circularity score is 0.

Assumptions & free parameters 4 free parameters · 5 assumptions · 1 invented entities

The framework introduces no new physical entities, but it depends on several domain assumptions: availability of skin masks for the ROI loss, transferability of ISIC 2019 pretraining, high-frequency gating as a useful inductive bias, and TTA+ensemble entropy as a useful uncertainty signal. The hand-set loss weights, augmentation counts, and the unreported HFSG cutoff are the main fitted/selected parameters. The HFSG module is an architectural device with no independent evidence beyond the paper's own ablation.

free parameters (4)
  • HFSG high-pass mask cutoff = not reported
    M_high in Eq. (2) requires a frequency cutoff/radius choice; the paper does not specify how it was set or whether it was tuned. This determines which high-frequency components are amplified.
  • Loss weights λ1-λ4 = 0.2 / 0.8 / 0.1 / 0.3
    Set by hand in Section 2.3/3.2 with no sensitivity analysis; they control the balance between masked focal, masked Dice, background penalty, and skin auxiliary losses.
  • K (ensemble folds) and N (TTA transforms) = K=5, N=8
    Arbitrary but standard choices for the uncertainty pipeline; no ablation studies their effect on the reported uncertainty maps or performance.
  • Number of artificial negative patches = 198
    The development set includes 198 negative patches to force healthy-skin discrimination; the choice of 198 and its effect on performance are not analyzed.
assumptions (5)
  • ad hoc to paper Pixel-level skin masks (or a skin segmentation head) are available for every clinical training image, as required by L_skin_aux and M_skin in Eq. (4)-(5)
    The paper never states where skin ground truth or the dynamic skin mask comes from; without it the ROI-constrained dual-task loss cannot be implemented as written.
  • domain assumption ISIC 2019 pretraining provides transferable dermatological features for clinical vitiligo photography
    Assumed in Section 2.2.1; the large measured jump from M1 to M2 supports it empirically, but no feature-level analysis is given.
  • domain assumption High-frequency gating improves vitiligo boundary delineation without amplifying noise
    Core hypothesis of HFSG; only the internal ablation (M4 vs M5) supports it, and the mechanism is not independently validated.
  • domain assumption TTA + ensemble entropy is a clinically meaningful uncertainty estimate
    The uncertainty maps are presented descriptively; no calibration, reliability diagram, or clinician study validates that high entropy corresponds to clinically ambiguous regions.
  • standard math 2D rFFT/IFFT and Fourier-domain operations as implemented in PyTorch behave as described
    Eqs. (1)-(3) rely on standard FFT mathematics and software implementations.
invented entities (1)
  • High-Frequency Spectral Gating (HFSG) module
    purpose: Explicitly filters high-frequency spectral components via FFT and injects them into spatial features to recover boundary textures.
    The module is evaluated only on the authors' own ablation and test set; no external benchmark, independent implementation, or analysis shows that the FFT-gating mechanism itself, rather than extra capacity, causes the gain.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Uncertainty-Aware Domain Adaptation for Vitiligo Segmentation in Clinical Photographs." pith.science (2026). https://pith.science/paper/BI5FFSMH

@misc{pith2026251211791,
  author       = {Pith},
  title        = {Pith review of: Uncertainty-Aware Domain Adaptation for Vitiligo Segmentation in Clinical Photographs},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/BI5FFSMH}},
  note         = {Machine review of arXiv:2512.11791}
}
read the original abstract

Accurately quantifying vitiligo extent in routine clinical photographs is crucial for longitudinal monitoring of treatment response. We propose a trustworthy, frequency-aware segmentation framework built on three synergistic pillars: (1) a data-efficient training strategy combining domain-adaptive pre-training on the ISIC 2019 dataset with an ROI-constrained dual-task loss to suppress background noise; (2) an architectural refinement via a ConvNeXt V2-based encoder enhanced with a novel High-Frequency Spectral Gating (HFSG) module and stem-skip connections to capture subtle textures; and (3) a clinical trust mechanism employing K-fold ensemble and Test-Time Augmentation (TTA) to generate pixel-wise uncertainty maps. Extensive validation on an expert-annotated clinical cohort demonstrates superior performance, achieving a Dice score of 85.05% and significantly reducing boundary error (95% Hausdorff Distance improved from 44.79 px to 29.95 px), consistently outperforming strong CNN (ResNet-50 and UNet++) and Transformer (MiT-B5) baselines. Notably, our framework demonstrates high reliability with zero catastrophic failures and provides interpretable entropy maps to identify ambiguous regions for clinician review. Our approach suggests that the proposed framework establishes a robust and reliable standard for automated vitiligo assessment.

Figures

Figures reproduced from arXiv: 2512.11791 by the authors.

Figure 1
Figure 1. Visual Abstract. Overview of the proposed framework addressing boundary ambiguity and trustworthy assessment in vitiligo segmentation. exhibit significant inter-observer variability, impeding standardized longitudinal monitoring (Nugroho et al., 2013). This underscores a well-recognized need for automated, objective measurement tools to support vitiligo management (Mazzetto et al., 2025). Automated segmentation of v… view at source ↗
Figure 2
Figure 2. Overview of the proposed Trustworthy Frequency-aware Segmenta￾tion Framework. The system is grounded on three synergistic pillars: (A) Domain-Adaptive Pre-training (Top-Left): The ConvNeXt V2 encoder is initialized with ISIC 2019 dermatological priors to mitigate domain shift. (B) High-Frequency Spectral Gating (Bottom-Left): A novel HFSG module explicitly filters spectral components via FFT to recover fine-grained … view at source ↗
Figure 3
Figure 3. Trustworthy Inference. Left: Segmentation result (Green: GT, Red: Pre￾diction, Dice: 0.8464). Middle: Predictive Entropy Map highlighting boundary ambiguity. Right: Variance Map showing model disagreement. fuzzy boundaries of depigmented lesions, our HFSG module successfully recovers high￾frequency boundary harmonics, reducing the Hausdorff Distance by over 30% compared to ResNet-50. Beyond quantitative metrics, our… view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Zero-shot Generalization on Dermoscopy. The model successfully segments lesions in dermoscopic images (unseen domain) despite significant differences in illumination and texture. 14 [PITH_FULL_IMAGE:figures/full_fig_p014_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

22 extracted references · 13 canonical work pages

  1. [1]

    Vitiligo: A Review

    Christina Bergqvist and Khaled Ezzedine. Vitiligo: A Review . Dermatology (Basel, Switzerland), 236 0 (6): 0 571--592, 2020. ISSN 1421-9832. doi:10.1159/000506103

    work page doi:10.1159/000506103 2020

  2. [2]

    Kristen Bibeau, Kathleen Butler, Mingyue Wang, Konstantina Skaltsa, and Iltefat H. Hamzavi. Psychometric Evaluation of the Facial and Total Vitiligo Area Scoring Index Instruments in the TRuE - V Phase 3 Studies . Dermatology and Therapy, 14 0 (8): 0 2223--2234, August 2024. ISSN 2193-8210. doi:10.1007/s13555-024-01223-y

    work page doi:10.1007/s13555-024-01223-y 2024

  3. [3]

    Faruk Abdullah Al Sohan, Rifat Al Mamun Rudro, and Shanaj Parvin

    Rahul Biswas, Md. Faruk Abdullah Al Sohan, Rifat Al Mamun Rudro, and Shanaj Parvin. VGG19 - ResSE : an optimized hybrid model for accurate segmentation and classification of vitiligo lesions. Neural Computing and Applications, 37 0 (32): 0 27227--27248, November 2025. ISSN 1433-3058. doi:10.1007/s00521-025-11670-z. URL https://doi.org/10.1007/s00521-025-11670-z

    work page doi:10.1007/s00521-025-11670-z 2025

  4. [4]

    Deep Learning - Based Segmentation of Lesions from Wide - Field Vitiligo Images

    Zhuangzhuang Fan and Changqing Wang. Deep Learning - Based Segmentation of Lesions from Wide - Field Vitiligo Images . International Journal of Pattern Recognition and Artificial Intelligence, 38 0 (10): 0 2457007, August 2024. ISSN 0218-0014. doi:10.1142/S0218001424570076. URL https://www.worldscientific.com/doi/10.1142/S0218001424570076. Publisher: Worl...

    work page doi:10.1142/s0218001424570076 2024

  5. [5]

    Maron, Julia Niebling, and Titus J

    Katharina Fogelberg, Sireesha Chamarthi, Roman C. Maron, Julia Niebling, and Titus J. Brinker. Domain shifts in dermoscopic skin cancer datasets: Evaluation of essential limitations for clinical translation. New Biotechnology, 76: 0 106--117, September 2023. ISSN 1871-6784. doi:10.1016/j.nbt.2023.04.006. URL https://www.sciencedirect.com/science/article/p...

    work page doi:10.1016/j.nbt.2023.04.006 2023

  6. [6]

    Adversarial Training Based Domain Adaptation of Skin Cancer Images

    Syed Qasim Gilani, Muhammad Umair, Maryam Naqvi, Oge Marques, and Hee-Cheol Kim. Adversarial Training Based Domain Adaptation of Skin Cancer Images . Life, 14 0 (8): 0 1009, August 2024. ISSN 2075-1729. doi:10.3390/life14081009. URL https://pmc.ncbi.nlm.nih.gov/articles/PMC11355601/

    work page doi:10.3390/life14081009 2024

  7. [7]

    Parametric modeling of narrowband UV - B phototherapy for vitiligo using a novel quantitative tool: the Vitiligo Area Scoring Index

    Iltefat Hamzavi, Hem Jain, David McLean, Jerry Shapiro, Haishan Zeng, and Harvey Lui. Parametric modeling of narrowband UV - B phototherapy for vitiligo using a novel quantitative tool: the Vitiligo Area Scoring Index . Archives of Dermatology, 140 0 (6): 0 677--683, June 2004. ISSN 0003-987X. doi:10.1001/archderm.140.6.677

    work page doi:10.1001/archderm.140.6.677 2004

  8. [8]

    WA - NET : enhanced boundary-aware segmentation of skin lesions via frequency-spatial feature fusion and attention-guided edge refinement

    Gefeng Hu, Wen Zhu, Xinyi Liao, and Qingbo Li. WA - NET : enhanced boundary-aware segmentation of skin lesions via frequency-spatial feature fusion and attention-guided edge refinement. Scientific Reports, 15 0 (1): 0 41598, November 2025. ISSN 2045-2322. doi:10.1038/s41598-025-25583-3. URL https://www.nature.com/articles/s41598-025-25583-3. Publisher: Na...

    work page doi:10.1038/s41598-025-25583-3 2025

Show all 22 references
  1. [9]

    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. Segment anything, 2023. URL https://arxiv.org/abs/2304.02643

    2023 arXiv

  2. [10]

    Evaluation of uncertainty estimation methods in medical image segmentation: Exploring the usage of uncertainty in clinical deployment

    Shiman Li, Mingzhi Yuan, Xiaokun Dai, and Chenxi Zhang. Evaluation of uncertainty estimation methods in medical image segmentation: Exploring the usage of uncertainty in clinical deployment. Computerized Medical Imaging and Graphics: The Official Journal of the Computerized Me...

    2025

  3. [11]

    Automating Vitiligo Skin Lesion Segmentation Using Convolutional Neural Networks

    Makena Low, Victor Huang, and Priyanka Raina. Automating Vitiligo Skin Lesion Segmentation Using Convolutional Neural Networks . In 2020 IEEE 17th International Symposium on Biomedical Imaging ( ISBI ) , pages 1--4, April 2020. doi:10.1109/ISBI45749.2020.9098682. URL https://i...

    2020

  4. [12]

    Potential of automated image analysis for the measurement of vitiligo lesions

    Roberto Mazzetto, Alvise Sernicola, Jacopo Tartaglia, Christian Ciolfi, and Mauro Alaibac. Potential of automated image analysis for the measurement of vitiligo lesions. Frontiers in Medicine, 12, August 2025. ISSN 2296-858X. doi:10.3389/fmed.2025.1623408. URL https://www.fron...

    2025

  5. [13]

    Semi- Automatic Tool for Vitiligo Detection and Analysis

    Paolo Neri, Michela Fiaschi, and Giovanni Menchini. Semi- Automatic Tool for Vitiligo Detection and Analysis . Journal of Imaging, 6 0 (3): 0 14, March 2020. ISSN 2313-433X. doi:10.3390/jimaging6030014. URL https://www.mdpi.com/2313-433X/6/3/14. Publisher: Multidisciplinary Di...

    2020 doi

  6. [14]

    Hani Ahmad Fadzil, Norashikin Shamsudin, and S

    Hermawan Nugroho, M. Hani Ahmad Fadzil, Norashikin Shamsudin, and S. H. Hussein. Computerised image analysis of vitiligo lesion: evaluation using manually defined lesion areas. Skin research and technology: official journal of International Society for Bioengineering and the S...

    2013

  7. [15]

    Mashagba, Azlan B

    Hamza Abu Owida, Ibrahim Abd El-Fattah, Suhaila Abuowaida, Nawaf Alshdaifat, Hamza A. Mashagba, Azlan B. Abd Aziz, Alaa Alzoubi, Samia Larguech, and Samir Salem Al-Bawri. A deep learning-based dual-branch framework for automated skin lesion segmentation and classification via ...

    2025 doi

  8. [16]

    Picardo, R

    M. Picardo, R. H. Huggins, H. Jones, R. Marino, M. Ogunsola, and J. Seneschal. The humanistic burden of vitiligo: a systematic literature review of quality-of-life outcomes. Journal of the European Academy of Dermatology and Venereology: JEADV, 36 0 (9): 0 1507--1523, Septembe...

    2022 doi

  9. [17]

    A frequency selection network for medical image segmentation

    Shu Tang, Haiheng Ran, Shuli Yang, Zhaoxia Wang, Wei Li, Haorong Li, and Zihao Meng. A frequency selection network for medical image segmentation. Heliyon, 10 0 (16): 0 e35698, August 2024. ISSN 2405-8440. doi:10.1016/j.heliyon.2024.e35698. URL https://www.sciencedirect.com/sc...

    2024 doi

  10. [18]

    SKIN - ORBIT : A bio-mimetic oscillatory resonance-based inference topology for universal skin lesion segmentation

    Anjali Thachankattil and Abhishek Sujith. SKIN - ORBIT : A bio-mimetic oscillatory resonance-based inference topology for universal skin lesion segmentation. Alexandria Engineering Journal, 128: 0 1177--1202, September 2025. ISSN 1110-0168. doi:10.1016/j.aej.2025.08.012. URL h...

    2025 doi

  11. [19]

    Decoding uncertainty for clinical decision-making

    Krasimira Tsaneva-Atanasova, Giulia Pederzanil, and Marianna Laviola. Decoding uncertainty for clinical decision-making. Philosophical transactions. Series A, Mathematical, physical, and engineering sciences, 383 0 (2292): 0 20240207, 2025. ISSN 1364-503X. doi:10.1098/rsta.202...

    2025

  12. [20]

    Aleatoric uncertainty estimation with test-time augmentation for medical image segmentation with convolutional neural networks

    Guotai Wang, Wenqi Li, Michael Aertsen, Jan Deprest, Sébastien Ourselin, and Tom Vercauteren. Aleatoric uncertainty estimation with test-time augmentation for medical image segmentation with convolutional neural networks. Neurocomputing, 338: 0 34--45, April 2019. ISSN 0925-23...

    2019 doi

  13. [21]

    Design and Assessment of Convolutional Neural Network Based Methods for Vitiligo Diagnosis

    Li Zhang, Suraj Mishra, Tianyu Zhang, Yue Zhang, Duo Zhang, Yalin Lv, Mingsong Lv, Nan Guan, Xiaobo Sharon Hu, Danny Ziyi Chen, and Xiuping Han. Design and Assessment of Convolutional Neural Network Based Methods for Vitiligo Diagnosis . Frontiers in Medicine, 8, October 2021....

    2021

  14. [22]

    MAFF - Net : SAM -powered Mixed Multi -scale Perception Adaptation and frequency-guided feature fusion for robust skin lesion segmentation

    Yifeng Zhou, Huiling Gong, Ruiyun Qiu, Shaofeng Wei, Zhixun Li, and Wei Zhang. MAFF - Net : SAM -powered Mixed Multi -scale Perception Adaptation and frequency-guided feature fusion for robust skin lesion segmentation. Biomedical Signal Processing and Control, 113: 0 109222, M...

    2026

Pith tools

Reviewed August 3, 2026 · model on record in the stance chip above.