Intelligent detection of areas requiring cosmetic treatment and simulation of treatment results using Convolutional Neural Networks and Generative Adversarial Networks

Document Type : Original Article

Authors

1 M.Sc. Student, Department of Mechatronics Engineering, SR.C., Islamic Azad University, Tehran, Iran

2 Associate Professor, Department of Mechanical Engineering, SR.C., Islamic Azad University, Tehran, Iran and Associate Professor, Modern Automotive Research Center, SR.C., Islamic Azad University, Tehran, Iran

Abstract

Nowadays, the demand for precise skin condition analysis and for predicting cosmetic treatment outcomes prior to intervention has become a critical requirement in dermatology and aesthetic medicine. In this study, we developed an intelligent deep learning–based framework for detecting facial areas requiring cosmetic intervention and for simulating post-treatment outcomes. To this end, a Convolutional Neural Network (CNN) was first employed for the automated identification of abnormalities such as wrinkles, pigmentation, and asymmetry. The proposed framework was implemented in Python using the PyTorch library and trained on a comprehensive dataset comprising 5,000 facial images from validated databases (CelebA, FFHQ, DermNet) spanning various age groups and skin types with rigorous manual annotations by three independent dermatology specialists for affected regions. Model evaluation was conducted based on quantitative metrics, including accuracy (94.2%), F1-Score (93.1%), MAE (0.023), and PSNR (31.5 dB), alongside an expert satisfaction score of 4.6 out of 5. The results demonstrate that the combined GAN-CNN architecture not only achieves precise identification of areas in need of intervention but also offers highly realistic and interpretable visualizations of treatment outcomes with integrated clinical personalization parameters.

Keywords

Main Subjects

References

[1]     H. Ding, E. Zhang, F. Fang, X. Liu, H. Zheng, H. Yang, Y. Ge, Y. Yang, and T. Lin, "Automatic identification of benign pigmented skin lesions from clinical images using deep convolutional neural network," BMC Biotechnol., vol. 22, p. 28, 2022. [Online]. Available: https://doi.org/10.1186/s12896-022-00755-5.
[2]     E. Tokgöz and M. A. Carro, "Applications of artificial intelligence, machine learning, and deep learning on facial plastic surgeries," in Cosmetic and Reconstructive Facial Plastic Surgery: A Review of Medical and Biomedical Engineering and Science Concepts, Springer, 2023.
[3]     Galić, M. Habijan, H. Leventić, and K. Romić, "Machine learning empowering personalized medicine: A comprehensive review of medical image analysis methods," Electronics, vol. 12, p. 4411, 2023. [Online]. Available: https://doi.org/10.3390/electronics12214411.
[4]     M. A. Kassem, K. M. Hosny, R. Damaševičius, and M. M. Eltoukhy, "Machine learning and deep learning methods for skin lesion classification and diagnosis: A systematic review," Diagnostics, vol. 11, p. 1390, 2021, pp. 423-445. [Online]. Available: https://doi.org/10.3390/diagnostics11081390
[5]     S. S. Noronha, M. A. Mehta, D. Garg, K. Kotecha, and A. Abraham, "Deep learning-based dermatological condition detection: A systematic review with recent methods, datasets, challenges, and future directions," IEEE Access, vol. 11, pp. 140348-140381, 2023. DOI: 10.1109/ACCESS.2023.3339635.
[6]     S. Q. Gilani and O. Marques, "Skin lesion analysis using generative adversarial networks: A review," Multimedia Tools Appl., vol. 82, no. 19, pp. 30065-30106, 2023. [Online]. Available: https://doi.org/10.1007/s11042-022-14267-z.
[7]     Z. Ali, S. Naz, H. Zaffar, J. Choi, and Y. Kim, "An IoMT-based melanoma lesion segmentation using conditional generative adversarial networks," Sensors, vol. 23, no. 7, p. 3548, 2023. [Online]. Available: https://doi.org/10.3390/s23073548.
[8]     J. F. Santos, L. del R. Silva-Calpa, F. G. de Souza, and K. Pal, "'Central Countries' and Brazil's Contributions to Nanotechnology," Curr. Nanomater., vol. 9, pp. 109-147, 2024. [Online]. Available: https://doi.org/10.2174/2405461508666230522564138.
[9]     L. Zang, X. Zhang, and B. Guo, "Federated deep reinforcement learning for online task offloading and resource allocation in WPC-MEC networks," IEEE Access, vol. 10, pp. 9856-9867, 2022. DOI: 10.1109/ACCESS.2022.3144415.
[10] A. Alshardan, S. Alahmari, M. Alghamdi, M. AL Sadig, A. Mohamed, and G. P. Mohammed, "GAN-based synthetic medical image augmentation for class imbalanced dermoscopic image analysis," Fractals, vol. 33, p. 2540039, 2025. [CrossRef].
[11] K. Behara, E. Bhero, and J. T. Agee, "Skin Lesion Synthesis and Classification Using an Improved DCGAN Classifier," Diagnostics, vol. 13, p. 2635, 2023. [CrossRef] [PubMed].
[12] E. Goceri, "GAN-based augmentation using a hybrid loss function for dermoscopy images," Artif. Intell. Rev., vol. 57, p. 234, 2024. [CrossRef].
[13] S. Innani, P. Dutande, U. Baid, V. Pokuri, S. Bakas, S. Talbar, B. Baheti, and S. C. Guntuku, "Generative adversarial networks based skin lesion segmentation," Sci. Rep., vol. 13, p. 13467, 2023. [CrossRef] [PubMed].
[14] A. Tavakkoli, "Generative Adversarial Networks in Medicine: Important Considerations for this Emerging Innovation in Artificial Intelligence," Ann. Biomed. Eng., vol. 51, pp. 2130–2142, 2023. [CrossRef].
[15] E. Perez and S. Ventura, "Progressive growing of GANs for improving data augmentation and skin cancer diagnosis," Artif. Intell. Med., vol. 141, p. 102556, 2023. [CrossRef].
[16] M. Salvi, S. Bassoli, F. Molinari, M. Havaei, A. Depeursinge, "Generative models for color normalization in pathology and dermatology," Expert. Syst. Appl., vol. 245, p. 123105, 2024. [CrossRef].
[17] N. Veeramani and P. Jayaraman, "A promising AI-based super resolution image reconstruction technique for early diagnosis of skin cancer," Sci. Rep., vol. 15, p. 5084, 2025. [CrossRef] [PubMed].
[18] T. Karras, S. Laine, S. Aittala, J. Hellsten, J. Lehtinen,T.Aila,“Analyzing and Improving the Image Quality of StyleGAN,”
in Proc. IEEE/CVF Conf. Computer Vision and Pattern Recognition (CVPR),
Seattle, WA, USA, Jun. 2020, pp. 8107–8116,
[19] P. Isola, J.Yan Zhu, T. Zhou, A. A. Efros,"Image-to-Image Translation with Conditional 1611.07004v3
[20] H. Sangkeun Jung," Variants of GAN: CGAN, Pix2Pix,CycleGAN,2024, https://medium.com/@hugmanskj/variants-of-gan-cgan-pix2pix-cyclegan-f7869de53152.
[21] A. Ray, A. Gupta, A. Al, “Skin Lesion Classification With Deep Convolutional Neural Network: Process Development and Validation,” *JMIR Dermatology*, vol. 3, no. 1, p. e18438, May 2020. doi:10.2196/18438.
[22] S. Innani, P. Dutande, U. Baid, V. Pokuri, S. Bakas, S. Talbar, B. Baheti, S. Chandra Guntuku, ― Generative adversarial networks based skin lesion segmentation", 2023, https://doi.org/10.1038/s41598-023-39648-8.
Applied and basic Machine intelligence research
Volume 4, Issue 1
May 2026
Pages 57-76

Files

Share

How to cite

Statistics