Benchmarking Pseudo-Mask Generation Methods for ResNet34-U-Net Acne Segmentation
DOI:
https://doi.org/10.30871/jaic.v10i4.13568Keywords:
Acne Lesion Segmentation, Pseudo-Mask Generation, U-Net, Superpixel, Weakly Supervised SegmentationAbstract
Acne lesion segmentation is essential for automated dermatological image analysis. However, developing accurate deep learning segmentation models requires pixel-level annotations, which are costly and time-consuming to obtain. To address this limitation, pseudo-mask generation methods can be utilized to produce surrogate labels for training segmentation networks without manual annotation. This study presents a comparative evaluation of five pseudo-mask generation methods for acne lesion segmentation, namely Contour-Based, Superpixel (SLIC), K-Means, Weakly Supervised Semantic Segmentation (WSSS) based on Otsu Thresholding and Morphological Operations, and Pseudo-Mask-Based Generation. The generated pseudo masks were employed as supervisory labels to train a ResNet34-U-Net segmentation model under identical experimental settings. To improve the robustness of the training process, data augmentation was applied exclusively to the training dataset. Segmentation performance was quantitatively evaluated using Dice Score, Intersection over Union (IoU), Precision, Recall and Validation Loss, and qualitatively assessed through visual comparisons of the generated pseudo masks and predicted segmentation results. The experimental results demonstrate that pseudo-mask quality has a substantial impact on segmentation performance. Among the evaluated methods, Superpixel (SLIC) achieved the highest performance with a Dice Score of 0.898, an IoU of 0.815, a Precision of 0.889, a Recall of 0.908, and the lowest validation loss of 0.268 indicating superior lesion boundary preservation and region consistency. WSSS (Otsu + Morphology) also produced competitive results, whereas Pseudo-Mask-Based and Contour-Based methods yielded comparatively lower performance. These findings demonstrate that high-quality pseudo masks can provide effective supervision for ResNet34-U-Net training, offering an annotation-efficient approach for acne lesion segmentation and providing practical insights into selecting suitable pseudo-mask generation methods for dermatological image analysis.
Downloads
References
[1] Z. Zhu et al., “Global, regional and national burdens of acne vulgaris in adolescents and young adults aged 10–24 years from 1990 to 2021: a trend analysis,” Br. J. Dermatol., vol. 192, no. 2, pp. 228–237, Jan. 2025, doi: 10.1093/bjd/ljae352.
[2] A. Wianti, H. Heni, R. Setyowati, R. Nuraeni, and A. Setyawati, “Adolescents and Early Adults With Acne Vulgaris: a Scoping Review on Psychological and Psychosocial Impacts and Quality of Life,” J. Matern. Care Reprod. Heal., vol. 6, no. 3, pp. 194–208, 2024, doi: 10.36780/jmcrh.v6i3.12267.
[3] S. Minaee, Y. Y. Boykov, F. Porikli, A. J. Plaza, N. Kehtarnavaz, and D. Terzopoulos, “Image Segmentation Using Deep Learning: A Survey,” IEEE Trans. Pattern Anal. Mach. Intell., pp. 1–1, 2021, doi: 10.1109/TPAMI.2021.3059968.
[4] R. Wang, T. Lei, R. Cui, B. Zhang, H. Meng, and A. K. Nandi, “Medical image segmentation using deep learning: A survey,” IET Image Process., vol. 16, no. 5, pp. 1243–1267, 2022, doi: 10.1049/ipr2.12419.
[5] P. Hiskiawan, Y. M. Geasela, H. Heryanto, E. Stephanie, M. Ardianti, and F. A. Sukarno, “Trustworthy Data Science Framework for Non-Invasive Nutritional Screening Using Computer Vision,” in 2025 International Conference on Informatics, Multimedia, Cyber and Information System (ICIMCIS), IEEE, Dec. 2025, pp. 1743–1748. doi: 10.1109/ICIMCIS68501.2025.11327268.
[6] P. Hiskiawan, T. P. Wijayanti, H. Heryanto, S. Everlin, S. A. Yasodhara, and D. Alexander, “Mel-Frequency Cepstral Coefficients and Neural Networks for Indonesian Traditional Music Recognition,” in 2025 International Conference on Informatics, Multimedia, Cyber and Information System (ICIMCIS), IEEE, Dec. 2025, pp. 1707–1712. doi: 10.1109/ICIMCIS68501.2025.11327436.
[7] J. Li et al., “Development and validation of an artificial intelligence-powered acne grading system incorporating lesion identification,” Front. Med., vol. 10, no. October, pp. 1–8, 2023, doi: 10.3389/fmed.2023.1255704.
[8] V. Venugopal, M. K. Nath, J. Joseph, and M. V. Das, “A deep learning-based illumination transform for devignetting photographs of dermatological lesions,” Image Vis. Comput., vol. 142, p. 104909, Feb. 2024, doi: 10.1016/j.imavis.2024.104909.
[9] M. Benčević, M. Habijan, I. Galić, D. Babin, and A. Pižurica, “Understanding skin color bias in deep learning-based skin lesion segmentation,” Comput. Methods Programs Biomed., vol. 245, no. November 2023, 2024, doi: 10.1016/j.cmpb.2024.108044.
[10] Z. He, X. Li, Y. Chen, N. Lv, and Y. Cai, “Attention-based dual-path feature fusion network for automatic skin lesion segmentation,” BioData Min., vol. 16, no. 1, pp. 1–22, 2023, doi: 10.1186/s13040-023-00345-x.
[11] L. Ran, Y. Li, G. Liang, and Y. Zhang, “Pseudo Labeling Methods for Semi-Supervised Semantic Segmentation: A Review and Future Perspectives,” IEEE Trans. Circuits Syst. Video Technol., vol. 35, no. 4, pp. 3054–3080, Apr. 2025, doi: 10.1109/TCSVT.2024.3508768.
[12] Y. Yu et al., “Techniques and Challenges of Image Segmentation: A Review,” Electronics, vol. 12, no. 5, p. 1199, Mar. 2023, doi: 10.3390/electronics12051199.
[13] M. Lee, S. Lee, J. Lee, and H. Shim, “Saliency as Pseudo-Pixel Supervision for Weakly and Semi-Supervised Semantic Segmentation,” IEEE Trans. Pattern Anal. Mach. Intell., vol. 45, no. 10, pp. 12341–12357, Oct. 2023, doi: 10.1109/TPAMI.2023.3273592.
[14] J. Feng, X. Wang, T. Li, S. Ji, and W. Liu, “Weakly-supervised semantic segmentation via online pseudo-mask correcting,” Pattern Recognit. Lett., vol. 165, pp. 33–38, Jan. 2023, doi: 10.1016/j.patrec.2022.11.024.
[15] D. O. Traini, G. Palmisano, C. Guerriero, and K. Peris, “Artificial Intelligence in the Assessment and Grading of Acne Vulgaris: A Systematic Review,” J. Pers. Med., vol. 15, no. 6, pp. 1–13, 2025, doi: 10.3390/jpm15060238.
[16] K. Yojana and L. Thillai Rani, “OCT layer segmentation using U-NET semantic segmentation and RESNET34 encoder-decoder,” Meas. Sensors, vol. 29, no. June, p. 100817, 2023, doi: 10.1016/j.measen.2023.100817.
[17] M. Moncho-Santonja, S. Aparisi-Navarro, B. Defez, and G. Peris-Fajarnés, “Segmentation of Acne Vulgaris Images Techniques: A Comparative and Technical Study,” Appl. Sci., vol. 13, no. 10, 2023, doi: 10.3390/app13106157.
[18] C. Yuan, D. Zhao, and S. S. Agaian, “MUCM-Net: a Mamba powered UCM-Net for skin lesion segmentation,” Explor. Med., vol. 5, pp. 694–708, 2024, doi: 10.37349/emed.2024.00250.
[19] L. Cai, K. Hou, and S. Zhou, “Intelligent skin lesion segmentation using deformable attention Transformer U-Net with bidirectional attention mechanism in skin cancer images,” Ski. Res. Technol., vol. 30, no. 8, pp. 1–13, 2024, doi: 10.1111/srt.13783.
[20] N. Lama, J. Hagerty, A. Nambisan, R. J. Stanley, and W. Van Stoecker, “Skin Lesion Segmentation in Dermoscopic Images with Noisy Data,” J. Digit. Imaging, vol. 36, no. 4, pp. 1712–1722, Apr. 2023, doi: 10.1007/s10278-023-00819-8.
[21] M. Tamoor, A. Naseer, A. Khan, and K. Zafar, “Skin Lesion Segmentation Using an Ensemble of Different Image Processing Methods,” Diagnostics, vol. 13, no. 16, pp. 1–14, 2023, doi: 10.3390/diagnostics13162684.
[22] A. K. Nugroho, Afiahayati, M. E. Wibowo, and H. Soebono, “The Effect of Preprocessing on Skin Lesion Segmentation,” in 2024 8th International Conference on Information Technology (InCIT), IEEE, Nov. 2024, pp. 455–460. doi: 10.1109/InCIT63192.2024.10810559.
[23] S. Joseph and O. O. Olugbara, “Preprocessing Effects on Performance of Skin Lesion Saliency Segmentation,” Diagnostics, vol. 12, no. 2, 2022, doi: 10.3390/diagnostics12020344.
[24] S. Jardim, J. António, and C. Mora, “Graphical Image Region Extraction with K-Means Clustering and Watershed,” J. Imaging, vol. 8, no. 6, pp. 1–27, 2022, doi: 10.3390/jimaging8060163.
[25] K. Zhu, N. N. Xiong, and M. Lu, “A Survey of Weakly-supervised Semantic Segmentation,” in 2023 IEEE 9th Intl Conference on Big Data Security on Cloud (BigDataSecurity), IEEE Intl Conference on High Performance and Smart Computing, (HPSC) and IEEE Intl Conference on Intelligent Data and Security (IDS), IEEE, May 2023, pp. 10–15. doi: 10.1109/BigDataSecurity-HPSC-IDS58521.2023.00013.
[26] J. Peng and Y. Wang, “Medical image segmentation with limited supervision: A review of deep network models,” IEEE Access, vol. 9, pp. 36827–36851, 2021, doi: 10.1109/ACCESS.2021.3062380.
[27] J. Su, Z. Luo, S. Lian, D. Lin, and S. Li, “Mutual learning with reliable pseudo label for semi-supervised medical image segmentation,” Med. Image Anal., vol. 94, p. 103111, May 2024, doi: 10.1016/j.media.2024.103111.
[28] P. Hiskiawan, M. K. Sari, R. E. Siregar, N. A. Valencia, and T. P. Wijayanti, “A Deep Learning Data Fusion Approach for Prostate MRI Zonal Segmentation Using Dual-Channel U-Net with T2 and ADC Images,” in 2026 International Conference on Current Research in Artificial Intelligence and Data Science (ICCRAIDS), IEEE, Apr. 2026, pp. 1–7. doi: 10.1109/ICCRAIDS67816.2026.11519671.
[29] K. He, X. Zhang, S. Ren, and J. Sun, “Deep residual learning for image recognition,” Proc. IEEE Comput. Soc. Conf. Comput. Vis. Pattern Recognit., vol. 2016-Decem, pp. 770–778, 2016, doi: 10.1109/CVPR.2016.90.
[30] P. Hiskiawan, C.-C. Chen, and Z.-K. Ye, “Processing of electrical resistivity tomography data using convolutional neural network in ERT-NET architectures,” Arab. J. Geosci., vol. 16, no. 10, p. 581, Oct. 2023, doi: 10.1007/s12517-023-11690-w.
[31] H. Liu et al., “MEA-Net: multilayer edge attention network for medical image segmentation,” Sci. Rep., vol. 12, no. 1, pp. 1–15, 2022, doi: 10.1038/s41598-022-11852-y.
[32] Z. Wang et al., “Revisiting Evaluation Metrics for Semantic Segmentation: Optimization and Evaluation of Fine-grained Intersection over Union,” Adv. Neural Inf. Process. Syst., vol. 36, no. NeurIPS, pp. 1–82, 2023.
[33] M. Yeung, L. Rundo, Y. Nan, E. Sala, C. B. Schönlieb, and G. Yang, “Calibrating the Dice Loss to Handle Neural Network Overconfidence for Biomedical Image Segmentation,” J. Digit. Imaging, vol. 36, no. 2, pp. 739–752, 2023, doi: 10.1007/s10278-022-00735-3.
[34] P. Fränti and R. Mariescu-Istodor, “Soft precision and recall,” Pattern Recognit. Lett., vol. 167, pp. 115–121, 2023, doi: 10.1016/j.patrec.2023.02.005.
Downloads
Published
How to Cite
Issue
Section
License
Copyright (c) 2026 Nikita Amelia Valencia, Puguh Hiskiawan
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
Authors who publish with this journal agree to the following terms:
- Authors retain copyright and grant the journal right of first publication with the work simultaneously licensed under a Creative Commons Attribution License (Attribution-ShareAlike 4.0 International (CC BY-SA 4.0) ) that allows others to share the work with an acknowledgement of the work's authorship and initial publication in this journal.
- Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgement of its initial publication in this journal.
- Authors are permitted and encouraged to post their work online (e.g., in institutional repositories or on their website) prior to and during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work (See The Effect of Open Access).
