CoChat
Sign Up
Cited by 129Cutaneous Melanoma Detection and ManagementAI in cancer detectionAdvanced Neural Network Applications

A novel vision transformer model for skin cancer classification

Open access
Yang, Guang, Luo, Suhuai, Greer, Peter·2023·Neural Processing Letters
DOI 10.1007/s11063-023-11204-5
Open this paper in CoChat — free accountOpen PDF

Abstract

Abstract Skin cancer can be fatal if it is found to be malignant. Modern diagnosis of skin cancer heavily relies on visual inspection through clinical screening, dermoscopy, or histopathological examinations. However, due to similarity among cancer types, it is usually challenging to identify the type of skin cancer, especially at its early stages. Deep learning techniques have been developed over the last few years and have achieved success in helping to improve the accuracy of diagnosis and classification. However, the latest deep learning algorithms still do not provide ideal classification accuracy. To further improve the performance of classification accuracy, this paper presents a novel method of classifying skin cancer in clinical skin images. The method consists of four blocks. First, class rebalancing is applied to the images of seven skin cancer types for better classification performance. Second, an image is preprocessed by being split into patches of the same size and then flattened into a series of tokens. Third, a transformer encoder is used to process the flattened patches. The transformer encoder consists of N identical layers with each layer containing two sublayers. Sublayer one is a multihead self-attention unit, and sublayer two is a fully connected feed-forward network unit. For each of the two sublayers, a normalization operation is applied to its input, and a residual connection of its input and its output is calculated. Finally, a classification block is implemented after the transformer encoder. The block consists of a flattened layer and a dense layer with batch normalization. Transfer learning is implemented to build the whole network, where the ImageNet dataset is used to pretrain the network and the HAM10000 dataset is used to fine-tune the network. Experiments have shown that the method has achieved a classification accuracy of 94.1%, outperforming the current state-of-the-art model IRv2 with soft attention on the same training and testing datasets. On the Edinburgh DERMOFIT dataset also, the method has better performance compared with baseline models.

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1007/s11063-023-11204-5

Cites

35 of 35 references verified0 citation errors · 0 not checkable

Statistics

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

  • Most prior studies address class imbalance through conventional image-level augmentation, which increases data volume but does not introduce genuine semantic diversity.

    “Most prior studies address this issue through conventional image-level augmentation (7, 14, 15, 17, 18, 20, 21, 23), which increases data volume but does not introduce genuine semantic diversity.”

    From Interpretable multimodal fusion for skin lesion classification using dermoscopic images and patient metadata (2026) · cites this paper as [7]

    The Methodology describes using conventional image augmentation to oversample underrepresented skin-cancer classes, and Related Work gives examples of earlier studies using augmentation or rebalancing. The paper does not establish that most prior studies used this approach or say that it fails to introduce genuine semantic diversity. The passages in our copy have no page numbers. (LLM, conf=0.85)

    Experimental — accuracy not yet measured

  • Most prior studies relied primarily on the fully connected layer for final classification.

    “In the classification stage, most prior studies relied primarily on the fully connected layer for final classification (7, 14, 16–18, 20).”

    From Interpretable multimodal fusion for skin lesion classification using dermoscopic images and patient metadata (2026) · cites this paper as [7]

    The Classification block section compares the paper’s dense-layer classification block with the original ViT’s MLP head, supporting the narrower point that an earlier model used such a head. The paper does not establish that most prior studies relied primarily on a fully connected layer for final classification; no page number is available in our copy. (LLM, conf=0.72)

    Experimental — accuracy not yet measured

  • Only limited works provided visual interpretation using attention maps or Grad-CAM.

    “Only limited works provided visual interpretation using attention maps (7) or Grad-CAM (13, 23, 28).”

    From Interpretable multimodal fusion for skin lesion classification using dermoscopic images and patient metadata (2026) · cites this paper as [7]

  • Class imbalance remains a central challenge in seven-class skin lesion classification using HAM10000.

    “Class imbalance remains a central challenge in seven-class skin lesion classification using HAM10000.”

    From Interpretable multimodal fusion for skin lesion classification using dermoscopic images and patient metadata (2026) · cites this paper as [7]

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Cites

Flagged first, then by citations

Research works printed but not linked: 6

  1. Bray F, Ferlay J, Soerjomataram I, Siegel RL, Torre LA, Jemal A (2018) Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA 68(6):394-424
  2. Siegel RL, Miller KD, Jemal A (2019) Cancer statistics. CA 69(1):7-3
  3. Vestergaard ME, Macaskill PH, Holt PE, Menzies SW (2008) Dermoscopy compared with naked eye examination for the diagnosis of primary melanoma: a meta-analysis of studies performed in a clinical setting. Br J Dermatol 159(3):669-676
  4. Datta SK, Shaikh MA, Srihari SN (2021) Soft Attention Improves Skin Cancer Classification Performance. InInterpretability of Machine Intelligence in Medical Image Computing, and Topological Data Analysis and Its Applications for Medical Data. Springer, Cham, pp 13-23
  5. Radford A, Wu J, Child R, Luan D, Amodei D, Sutskever I (2019) Language models are unsupervised multitask learners. OpenAI blog 1(8):9
  6. Tolstikhin IO, Houlsby N, Kolesnikov A, Beyer L, Zhai X, Unterthiner T, Yung J, Steiner A, Keysers D, Uszkoreit J, Lucic M (2021) Mlp-mixer: an all-mlp architecture for vision. Adv Neural Inf Process Syst 6(34):24261-24272

Cited by

Sorted by citations

Related

Authors
Yang, G., Luo, S., & Greer, P.
Affiliations
University of Newcastle Australia (AU)
Corresponding author
Yang, Guang
ORCID
Venue
Neural Processing Letters
Publisher
Springer Science and Business Media LLC
Published
2023-03-27
Volume
55
Issue
7
Pages
9335-9351
ISSN
  • 1370-4621 (ISSN-L)
  • 1573-773X
DOI
10.1007/s11063-023-11204-5
OpenAlex
W4360994757
Cited by
129
References
31
Citations by year
2026: 48 · 2025: 43 · 2024: 32 · 2023: 6
FWCI
15.77
Type
journal article
Language
en
Open access
Licence
Locations
  • Neural Processing Letters (publishedVersion, cc-by)
  • Figshare (submittedVersion, cc-by)
  • NOVA (University of Newcastle Australia) (submittedVersion, cc-by)
Funders
The University of Newcastle
Topics
  • Health Sciences › Medicine › Oncology › Cutaneous Melanoma Detection and Management
  • Physical Sciences › Computer Science › Artificial Intelligence › AI in cancer detection
  • Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Advanced Neural Network Applications
Keywords
skin cancer classification, vision transformer, Transformer encoder, multi-head self-attention, class rebalancing, dermoscopic images, transfer learning, HAM10000 dataset

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review