A Data-Driven Closed-Loop Parameter Adaptive Framework for Automated Visual Standardization Systems
DOI:
https://doi.org/10.54097/6mqzwy06Keywords:
Affine transformation, Closed-loop control, Edge integrity, Parameter adaptation, Portrait segmentation, Visual standardizationAbstract
Automated visual standardization — the conversion of arbitrary user-provided portraits into images that satisfy strict format specifications — is a key component of high-reliability identity-management systems. Conventional pipelines operate as feedforward open-loop systems: they map facial landmarks to a fixed scaling ratio and synthesize the output in a single pass, implicitly assuming a linear relationship between facial scale and body anatomy. Under high-dimensional stochastic geometric variation (narrow or wide shoulders, long hair, atypical clothing), this assumption breaks down and produces shoulder truncation or bottom-edge gaps that violate format compliance. This paper reformulates visual standardization as a feedback-controlled parameter-optimization process. Our framework cascades perception-aware color normalization, BiRefNet-based foreground segmentation with distance-transform edge purification, and a closed-loop geometric adaptation stage in which an edge-integrity sensor measures posterior canvas coverage and a controller iteratively refines the affine scale parameter until convergence. On 120 real-world samples spanning 12 challenging categories, the closed-loop framework raises the bottom-edge touch rate from 99.17% to 100% and bottom-row foreground coverage from 98.73% to 99.99%, while reducing the boundary contamination ratio from 67.5% to 53.8%. Ablation studies confirm that the geometric feedback loop is the sole factor eliminating bottom-edge gaps and that edge purification independently accounts for a 15.9-percentage-point reduction in boundary contamination. The results demonstrate that posterior feedback converts residual geometric uncertainty into a convergent control problem, enabling deterministic compliance under strict business constraints.
Downloads
References
[1] Zheng, P., Xu, J., Liu, Z., Jiang, Y., Chen, Y., & Fan, H. (2024). Bilateral reference for high-resolution dichotomous image segmentation. CAAI Artificial Intelligence Research, 3, Article 9150038. https://doi.org/10.26599/AIR.2024.9150038 DOI: https://doi.org/10.26599/AIR.2024.9150038
[2] Qin, X., Zhang, Z., Huang, C., Dehghan, M., Zaiane, O. R., & Jagersand, M. (2020). U2-Net: Going deeper with nested U-structure for salient object detection. Pattern Recognition, 106, Article 107404. https://doi.org/10.1016/j.patcog.2020.107404 DOI: https://doi.org/10.1016/j.patcog.2020.107404
[3] Ke, Z., Sun, J., Li, K., Yan, Q., & Lau, R. W. H. (2022). MODNet: Real-time trimap-free portrait matting via objective decomposition. In Proceedings of the AAAI Conference on Artificial Intelligence (Vol. 36, No. 1, pp. 1140-1147). AAAI Press. https://doi.org/10.1609/aaai.v36i1.19998 DOI: https://doi.org/10.1609/aaai.v36i1.19999
[4] Porter, T., & Duff, T. (1984). Compositing digital images. ACM SIGGRAPH Computer Graphics, 18(3), 253-259. https://doi.org/10.1145/964965.808606 DOI: https://doi.org/10.1145/964965.808606
[5] Levin, A., Lischinski, D., & Weiss, Y. (2008). A closed-form solution to natural image matting. IEEE Transactions on Pattern Analysis and Machine Intelligence, 30(2), 228-242. https://doi.org/10.1109/TPAMI.2007.1177 DOI: https://doi.org/10.1109/TPAMI.2007.1177
[6] Rosenfeld, A., & Pfaltz, J. L. (1966). Sequential operations in digital picture processing. Journal of the ACM, 13(4), 471-494. https://doi.org/10.1145/321356.321357 DOI: https://doi.org/10.1145/321356.321357
[7] Felzenszwalb, P. F., & Huttenlocher, D. P. (2012). Distance transforms of sampled functions. Theory of Computing, 8(1), 415-428. https://doi.org/10.4086/toc.2012.v008a019 DOI: https://doi.org/10.4086/toc.2012.v008a019
[8] Åström, K. J., & Murray, R. M. (2008). Feedback systems: An introduction for scientists and engineers. Princeton University Press. DOI: https://doi.org/10.1515/9781400828739
[9] Wolberg, G. (1990). Digital image warping. IEEE Computer Society Press.
[10] Viola, P., & Jones, M. (2001). Rapid object detection using a boosted cascade of simple features. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR) (Vol. 1, pp. 511-518). IEEE. https://doi.org/10.1109/CVPR.2001.990517 DOI: https://doi.org/10.1109/CVPR.2001.990517
[11] Land, E. H. (1977). The Retinex theory of color vision. Scientific American, 237(6), 108-128. https://doi.org/10.1038/scientificamerican1277-108 DOI: https://doi.org/10.1038/scientificamerican1277-108
[12] Jobson, D. J., Rahman, Z., & Woodell, G. A. (1997). A multiscale retinex for bridging the gap between color images and the human observation of scenes. IEEE Transactions on Image Processing, 6(7), 965-976. https://doi.org/10.1109/83.597272 DOI: https://doi.org/10.1109/83.597272
[13] Guo, X., Li, Y., & Ling, H. (2017). LIME: Low-light image enhancement via illumination map estimation. IEEE Transactions on Image Processing, 26(2), 982-993. https://doi.org/10.1109/TIP.2016.2639450 DOI: https://doi.org/10.1109/TIP.2016.2639450
[14] Guo, C., Li, C., Guo, J., Loy, C. C., Hou, J., Kwong, S., & Cong, R. (2020). Zero-reference deep curve estimation for low-light image enhancement. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (pp. 1780-1789). IEEE. https://doi.org/10.1109/CVPR42600.2020.00185 DOI: https://doi.org/10.1109/CVPR42600.2020.00185
[15] Hu, G., & Jamaluddin, R. A. (2026). An adaptive portrait image enhancement algorithm based on facial and scene statistics in complex lighting. Manuscript submitted for publication.
Downloads
Published
Issue
Section
License
Copyright (c) 2026 Journal of Computing and Electronic Information Management
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
