S-SYNTH: Knowledge-Based, Synthetic Generation of Skin Images (2408.00191v1)
Abstract: Development of AI techniques in medical imaging requires access to large-scale and diverse datasets for training and evaluation. In dermatology, obtaining such datasets remains challenging due to significant variations in patient populations, illumination conditions, and acquisition system characteristics. In this work, we propose S-SYNTH, the first knowledge-based, adaptable open-source skin simulation framework to rapidly generate synthetic skin, 3D models and digitally rendered images, using an anatomically inspired multi-layer, multi-component skin and growing lesion model. The skin model allows for controlled variation in skin appearance, such as skin color, presence of hair, lesion shape, and blood fraction among other parameters. We use this framework to study the effect of possible variations on the development and evaluation of AI models for skin lesion segmentation, and show that results obtained using synthetic data follow similar comparative trends as real dermatologic images, while mitigating biases and limitations from existing datasets including small dataset size, lack of diversity, and underrepresentation.
- Evaluation of diagnosis diversity in artificial intelligence datasets: a scoping review. British Journal of Dermatology, 2023.
- Towards reliable dermatology evaluation benchmarks. In Machine Learning for Health (ML4H). PMLR, 2023.
- Estimating skin tone and effects on classification performance in dermatology datasets. arXiv:1910.13268, 2019.
- Understanding skin color bias in deep learning-based skin lesion segmentation. Computer Methods and Programs in Biomedicine, 2024.
- The stochastic digital human is now enrolling for in silico imaging trials–methods and tools for generating digital cohorts. Progress in Biomedical Engineering, 2023.
- A survey on deep learning for skin lesion segmentation. Medical Image Analysis, 2023.
- A skin lesion hair mask dataset with fine-grained annotations. Data in Brief, 2023.
- Lack of transparency and potential bias in artificial intelligence data sets and algorithms: a scoping review. JAMA Dermatology, 2021.
- Melanogans: high resolution skin lesion synthesis with gans. arXiv:1804.04338, 2018.
- Skin lesion synthesis and classification using an improved dcgan classifier. Diagnostics, 2023.
- Dermgan: Synthetic generation of clinical skin images with pathology. In Machine Learning for Health (ML4H). PMLR, 2020.
- Controlled synthesis of dermoscopic images via a new color labeled generative style transfer network to enhance melanoma segmentation. In International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC). IEEE, 2018.
- Dario Augusto Borges Oliveira. Controllable skin lesion synthesis using texture patches, bézier curves and conditional gans. In International Symposium on Biomedical Imaging (ISBI). IEEE, 2020.
- Improving skin color diversity in cancer detection: deep learning approach. JMIR Dermatology, 2022.
- Improving dermatology classifiers across populations using images generated by large diffusion models. arXiv:2211.13352, 2022.
- Augmenting medical image classifiers with synthetic data from latent diffusion models. arXiv:2308.12453, 2023.
- Sidefx: Houdini, 2022. https://www.sidefx.com/docs/houdini/index.html.
- Mitsuba 3 renderer, 2022. https://mitsuba-renderer.org.
- The structure and function of skin. In Dermatological and Transdermal Formulations. CRC press, 2002.
- Computational model of tumor growth for in silico trials. In Medical Imaging 2021: Physics of Medical Imaging. SPIE, 2021.
- Automated animation pipeline for visualizing in silico tumor growth models. In Medical Imaging 2023: Physics of Medical Imaging. SPIE, 2023.
- Steven L Jacques. Optical properties of biological tissues: a review. Physics in Medicine & Biology, 2013.
- Toward accurate quantitative photoacoustic imaging: learning vascular blood oxygen saturation in three dimensions. Journal of Biomedical Optics, 25(8):085003–085003, 2020.
- Microfacet models for refraction through rough surfaces. In Eurographics Conference on Rendering Techniques, 2007.
- Poly Haven: the public 3D asset library. https://polyhaven.com/hdris.
- Ege-unet: an efficient group enhanced unet for skin lesion segmentation. In International Conference on Medical Image Computing and Computer-Assisted Intervention (MICCAI). Springer, 2023.
- Dermosegdiff: A boundary-aware segmentation diffusion model for skin lesion delineation. In International Workshop on PRedictive Intelligence In MEdicine (PRIME). Springer, 2023.
- Skin lesion analysis toward melanoma detection 2018: A challenge hosted by the international skin imaging collaboration (ISIC). arXiv:1902.03368, 2019.
- The HAM10000 dataset, a large collection of multi-source dermatoscopic images of common pigmented skin lesions. scientific data. 2018; 5: 180161. Search in, 2018.
- Transnorm: Transformer provides a strong spatial normalization mechanism for a deep segmentation model. IEEE Access, 2022.
- Real-valued (medical) time series generation with recurrent conditional gans. arXiv:1706.02633, 2017.
- Domain adaptation for medical image analysis: a survey. IEEE Transactions on Biomedical Engineering, 2021.
- Digital hair removal by deep learning for skin lesion segmentation. Pattern Recognition, 2021.
- Andrea Kim (1 paper)
- Niloufar Saharkhiz (3 papers)
- Elena Sizikova (19 papers)
- Miguel Lago (3 papers)
- Berkman Sahiner (13 papers)
- Jana Delfino (2 papers)
- Aldo Badano (5 papers)