Smart Visual Optimizer: An Adaptive Human-in-the-Loop System for Intelligent Image Enhancement and Quality-Aware Decision Support
DOI:
https://doi.org/10.65405/sgpsep53Keywords:
Image Enhancement, Image Restoration, Human-Computer Interaction, Human-in-the-Loop, Artificial Intelligence, Image Quality Assessment, Adaptive Image Processing, Computer VisionAbstract
Image enhancement remains a challenging problem due to the diversity of image degradation types, varying user preferences,
and different application requirements. Traditional enhancement methods often rely on fixed processing pipelines that may not
adapt effectively to different image characteristics, which can produce over-enhanced or visually inconsistent results.
This paper presents Smart Visual Optimizer, a human-centered adaptive image enhancement framework that combines scene-
aware analysis, mode-specific image processing pipelines, objective quality assessment, and interactive user control. The proposed
system analyzes image characteristics and supports five enhancement modes: AI Super, Product, Social, Low Light, and Scan. Each
mode applies a dedicated processing route designed for its target image category, such as product photographs, portrait images,
low-light scenes, general images, and scanned documents. The framework computes quality indicators derived from brightness,
contrast, sharpness, saturation, highlight distribution, and shadow distribution, then displays the enhanced result with quality metrics
and system recommendations.
Experimental evaluation was conducted using 25 real-world images across the supported categories. The results showed
consistent improvement in the average quality score across all tested categories, with the strongest improvements observed in Low
Light, Document, and General images. Product and Social modes applied more conservative enhancement to preserve realistic
appearance, natural colors, skin texture, and identity.
The proposed framework demonstrates that combining adaptive image processing with Human-in-the-Loop interaction can
improve practical image enhancement systems by providing automated optimization, visual comparison, quality-aware feedback,
and user-controlled refinement.
Downloads
References
[1] OpenCV, “Image thresholding,” https://docs.opencv.org/4.x/d7/d4d/tutorialpythresholding.html, accessed : 2026−06−17.
[2] ——, “Histogram equalization and clahe,” https://docs.opencv.org/4.x/d5/daf/tutorialpyhistogramequalization.html, 2026, accessed :
2026 −06 −17.
[3] ——, “Image denoising,” https://docs.opencv.org/4.x/d5/d69/tutorialpynonlocalmeans.html, 2026, accessed : 2026 −06 −17.
[4] ——, “Morphological transformations,” https://docs.opencv.org/4.x/d9/d61/tutorialpymorphologicalops.html, 2026, accessed :
2026 −06 −17.
[5] Y. Cai, H. Bian, J. Lin, H. Wang, R. Timofte, and Y. Zhang, “Retinexformer: One-stage retinex-based transformer for low-light
image enhancement,” in Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), 2023.
[6] S. W. Zamir, A. Arora, S. Khan, M. Hayat, F. S. Khan, and M.-H. Yang, “Restormer: Efficient transformer for high-resolution
image restoration,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2022,
pp. 5718–5729.
[7] Y. Ai, H. Huang, X. Zhou, J. Wang, and R. He, “Multimodal prompt perceiver: Empower adaptiveness, generalizability
and fidelity for all-in-one image restoration,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern
Recognition (CVPR), 2024, pp. 25 432–25 444.
[8] X. Wang, L. Xie, C. Dong, and Y. Shan, “Real-esrgan: Training real-world blind super-resolution with pure synthetic data,” in
Proceedings of the IEEE/CVF International Conference on Computer Vision Workshops (ICCVW), 2021, pp. 1905–1914.
[9] X. Wang, Y. Li, H. Zhang, and Y. Shan, “Towards real-world blind face restoration with generative facial prior,” in Proceedings
of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2021, pp. 9168–9178.
[10] K. Ma, Z. Shu, X. Bai, J. Wang, and D. Samaras, “Docunet: Document image unwarping via a stacked u-net,” in Proceedings
of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2018, pp. 4700–4709.
[11] I. Pratikakis, K. Zagoris, X. Karagiannis, L. Tsochatzidis, T. Mondal, and I. Marthot-Santaniello, “Icdar 2019 competition
on document image binarization (dibco 2019),” in Proceedings of the International Conference on Document Analysis and
Recognition (ICDAR), 2019, pp. 1547–1556.
[12] U.-V. Marti and H. Bunke, “The iam-database: An english sentence database for off-line handwriting recognition,” International
Journal on Document Analysis and Recognition, vol. 5, pp. 39–46, 2002.
[13] Z. Wang, A. C. Bovik, H. R. Sheikh, and E. P. Simoncelli, “Image quality assessment: From error visibility to structural
similarity,” IEEE Transactions on Image Processing, vol. 13, no. 4, pp. 600–612, 2004.
[14] B. Shneiderman, “Human-centered artificial intelligence: Reliable, safe and trustworthy,” International Journal of Human–
Computer Interaction, vol. 36, no. 6, pp. 495–504, 2020.
[15] N. Zheng, M. Zhou, Y. Dong, X. Rui, J. Huang, C. Li, and F. Zhao, “Empowering low-light image enhancer through customized
learnable priors,” in Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), 2023, pp. 12 525–12 535.
[16] C. Tian, Y. Xu, and W. Zuo, “A survey of deep learning-based low-light image enhancement,” Sensors, vol. 23, no. 7, p. 3567,
2023.
[17] A. Mittal, A. K. Moorthy, and A. C. Bovik, “No-reference image quality assessment in the spatial domain,” IEEE Transactions
on Image Processing, vol. 21, no. 12, pp. 4695–4708, 2012.
