Llorens Quintana, ClaraKuran, UmutKuran, Emre CanMadrid Costa, David2025-11-142025-11-142025-09Llorens-Quintana C, Kuran U, Kuran EC, Madrid-Costa D. Enhancing Conjunctival Vasculature Imaging: A Multi-Objective Cuckoo Search Approach for Contrast Enhancement Optimization. Transl Vis Sci Technol. 2025 Sep 2;14(9):36. doi: 10.1167/tvst.14.9.36. PMID: 41002106; PMCID: PMC12489868.10.1167/tvst.14.9.36https://hdl.handle.net/20.500.14352/126107Purpose: To develop and evaluate an automated method for enhancing the quality of vascular conjunctival images through optimized contrast and reduced noise. Methods: Conjunctival images were acquired using a functional slit lamp biomicroscope. The visibility of the vascular structures was enhanced using contrast limited adaptive histogram equalization (CLAHE). The multi-objective cuckoo search (MOCS) optimization algorithm was implemented to tune CLAHE hyperparameters with two objective functions that maximize image contrast and minimize noise amplification. All the images were enhanced using CLAHE with optimized parameters (MOCS-CLAHE) and with predetermined parameters (CLAHE). The performance of both approaches was evaluated through qualitative assessment and quantitative image quality metrics, including peak signal-to-noise ratio (PSNR), structural similarity index (SSIM), and natural image quality evaluator (NIQE). Results: Both approaches significantly increased the contrast of the original conjunctival images. Despite CLAHE generating images with higher vessel contrast than MOCS-CLAHE, it increases image noise significantly. The overall vessel visibility and quality of enhanced images was significantly better with MOCS-CLAHE, consistently giving higher PSNR and SSIM, and lower NIQE compared to CLAHE. Conclusions: MOCS optimization is an efficient method to estimate CLAHE parameters when preprocessing conjunctival images acquired with a slit lamp. It provides high quality images of the conjunctival vasculature emphasizing vessels structure by increasing contrast and keeping noise amplification to a minimum. Translational relevance: This automated enhancement technique may improve subsequent image segmentation and classification tasks in conjunctival imaging by optimizing contrast and noise reduction for each individual image, thus contributing to more reliable and efficient diagnostic procedures.engAttribution-ShareAlike 4.0 Internationalhttp://creativecommons.org/licenses/by-sa/4.0/journal articlehttps://doi.org/10.1167/tvst.14.9.3641002106open access617.7Ciencias Biomédicas32 Ciencias Médicas