Synthetic Melanoma Image Generation and Evaluation Using Generative Adversarial Networks

• URL: http://arxiv.org/abs/2603.13497v1
• Date: Fri, 13 Mar 2026 18:21:55 GMT
• Title: Synthetic Melanoma Image Generation and Evaluation Using Generative Adversarial Networks
• Abstract summary: We compare four GAN architectures-DCGAN, StyleGAN2, and two StyleGAN3 variants (T/R)-for high-resolution melanoma-specific synthesis.<n>StyleGAN2 achieves the best balance of quantitative performance and perceptual quality, attaining FID scores of 24.8 and 7.96 (ISIC and 7.96 (ISIC 2020) at gamma=0.8.<n>In a controlled augmentation experiment, adding synthetic melanoma images to address class imbalance improved melanoma detection AUC from 0.925 to 0.945 on a held-out real-image test set.
• Score: 8.789857856546147
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Melanoma is the most lethal form of skin cancer, and early detection is critical for improving patient outcomes. Although dermoscopy combined with deep learning has advanced automated skin-lesion analysis, progress is hindered by limited access to large, well-annotated datasets and by severe class imbalance, where melanoma images are substantially underrepresented. To address these challenges, we present the first systematic benchmarking study comparing four GAN architectures-DCGAN, StyleGAN2, and two StyleGAN3 variants (T/R)-for high-resolution melanoma-specific synthesis. We train and optimize all models on two expert-annotated benchmarks (ISIC 2018 and ISIC 2020) under unified preprocessing and hyperparameter exploration, with particular attention to R1 regularization tuning. Image quality is assessed through a multi-faceted protocol combining distribution-level metrics (FID), sample-level representativeness (FMD), qualitative dermoscopic inspection, downstream classification with a frozen EfficientNet-based melanoma detector, and independent evaluation by two board-certified dermatologists. StyleGAN2 achieves the best balance of quantitative performance and perceptual quality, attaining FID scores of 24.8 (ISIC 2018) and 7.96 (ISIC 2020) at gamma=0.8. The frozen classifier recognizes 83% of StyleGAN2-generated images as melanoma, while dermatologists distinguish synthetic from real images at only 66.5% accuracy (chance = 50%), with low inter-rater agreement (kappa = 0.17). In a controlled augmentation experiment, adding synthetic melanoma images to address class imbalance improved melanoma detection AUC from 0.925 to 0.945 on a held-out real-image test set. These findings demonstrate that StyleGAN2-generated melanoma images preserve diagnostically relevant features and can provide a measurable benefit for mitigating class imbalance in melanoma-focused machine learning pipelines.

Related papers

• BenchX: Benchmarking AI Models for Cancer Detection and Localization with Demographic and Protocol Biases [59.19934491064188]
  Artificial intelligence (AI) has achieved remarkable success in medical imaging.<n>These models often perform inconsistently across real-world clinical settings.<n>We develop a benchmark of 85,355 CT scans that systematically evaluates 12 tumor-detection AI models.
  arXiv  Detail & Related papers  (2026-06-23T17:58:59Z)
• A Generative AI Approach for Reducing Skin Tone Bias in Skin Cancer Classification [2.676349883103404]
  Skin cancer is one of the most common cancers worldwide and early detection is critical for effective treatment.<n>Current AI diagnostic tools are often trained on datasets dominated by lighter skin tones, leading to reduced accuracy and fairness for people with darker skin.<n>This paper addresses the challenge of skin tone imbalance in automated skin cancer detection using dermoscopic images.
  arXiv  Detail & Related papers  (2026-02-16T00:20:56Z)
• SkinGenBench: Generative Model and Preprocessing Effects for Synthetic Dermoscopic Augmentation in Melanoma Diagnosis [0.0]
  generative architecture choice has a stronger influence on both image fidelity and diagnostic utility than preprocessing complexity.<n>Synthetic data augmentation substantially improved melanoma detection with 8-15% absolute gains in melanoma F1-score.
  arXiv  Detail & Related papers  (2025-12-19T13:52:11Z)
• Comparing Conditional Diffusion Models for Synthesizing Contrast-Enhanced Breast MRI from Pre-Contrast Images [31.354231459573356]
  Dynamic contrast-enhanced (DCE) MRI is essential for breast cancer diagnosis and treatment.<n>However, its reliance on contrast agents introduces safety concerns, contraindications, increased cost, and workflow complexity.<n>We present pre-contrast conditioned denoising diffusion probabilistic models to synthesize DCE-MRI.
  arXiv  Detail & Related papers  (2025-08-19T12:24:55Z)
• HepatoGEN: Generating Hepatobiliary Phase MRI with Perceptual and Adversarial Models [33.7054351451505]
  We propose a deep learning based approach for synthesizing hepatobiliary phase (HBP) images from earlier contrast phases.<n> Quantitative evaluation using pixel-wise and perceptual metrics, combined with blinded radiologist reviews, showed that pGAN achieved the best quantitative performance.<n>In contrast, the U-Net produced consistent liver enhancement with fewer artifacts, while DDPM underperformed due to limited preservation of fine structural details.
  arXiv  Detail & Related papers  (2025-04-25T15:01:09Z)
• Robust Melanoma Thickness Prediction via Deep Transfer Learning enhanced by XAI Techniques [39.97900702763419]
  This study focuses on analyzing dermoscopy images to determine the depth of melanomas. The Breslow depth, measured from the top of the granular layer to the deepest point of tumor invasion, serves as a crucial parameter for staging melanoma and guiding treatment decisions. Various datasets, including ISIC and private collections, were used, comprising a total of 1162 images. Results indicated that the models achieved significant improvements over previous methods.
  arXiv  Detail & Related papers  (2024-06-19T11:07:55Z)
• Improving Breast Cancer Grade Prediction with Multiparametric MRI Created Using Optimized Synthetic Correlated Diffusion Imaging [71.91773485443125]
  Grading plays a vital role in breast cancer treatment planning. The current tumor grading method involves extracting tissue from patients, leading to stress, discomfort, and high medical costs. This paper examines using optimized CDI$s$ to improve breast cancer grade prediction.
  arXiv  Detail & Related papers  (2024-05-13T15:48:26Z)
• Using Whole Slide Image Representations from Self-Supervised Contrastive Learning for Melanoma Concordance Regression [2.21878843241715]
  Diagnosis is complicated by a high discordance rate among pathologists when distinguishing between melanoma and benign melanocytic lesions. We present a melanoma concordance regression deep learning model capable of predicting the concordance rate of invasive melanoma or melanoma in-situ from digitized Whole Slide Images (WSIs) We trained a SimCLR feature extractor with 83,356 WSI tiles randomly sampled from 10,895 specimens originating from four distinct pathology labs. We achieved a Root Mean Squared Error (RMSE) of 0.28 +/- 0.01 on the test set and a precision and recall of 0.85 +/- 0.05 and 0.61 +/- 0.06, respectively
  arXiv  Detail & Related papers  (2022-10-10T16:07:41Z)
• Extended Feature Space-Based Automatic Melanoma Detection System [9.165013127586267]
  Melanoma is the deadliest form of skin cancer. Uncontrollable growth of melanocytes leads to melanoma. The Automatic Melanoma Detection System (AMDS) helps to detect melanoma based on image processing techniques. A novel algorithm ExtFvAMDS is proposed for the calculation of Extended Feature Vector Space.
  arXiv  Detail & Related papers  (2022-09-10T04:15:45Z)
• Visualizing CoAtNet Predictions for Aiding Melanoma Detection [0.0]
  This paper proposes a multi-class classification task using the CoAtNet architecture. It achieves an overall precision of 0.901, recall 0.895, and AP 0.923, indicating high performance compared to other state-of-the-art networks.
  arXiv  Detail & Related papers  (2022-05-21T06:41:52Z)
• Melanoma Diagnosis with Spatio-Temporal Feature Learning on Sequential Dermoscopic Images [40.743870665742975]
  Existing dermatologists for automated melanoma diagnosis are based on single-time point images of lesions. We propose an automated framework for melanoma diagnosis using sequential dermoscopic images.
  arXiv  Detail & Related papers  (2020-06-19T04:08:22Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.