COMPUTER-BASED SIMULATION OF OCULAR PROSTHETIC REHABILITATION: A PATIENT-CENTERED TOOL TO ENHANCE AESTHETIC OUTCOMES

Abstract

Eye loss affects facial symmetry, self-image, social interaction, and quality of life; therefore, ocular prosthetic rehabilitation must address not only anatomical restoration, but also the patient’s expectations, confidence, and understanding of the treatment process. The aim of this project was to develop and evaluate a computer-based simulation tool capable of generating a preliminary aesthetic visualization of ocular prosthetic rehabilitation in patients with unilateral eye loss. A mixed experimental approach was implemented in three sequential phases: (1) collection of functional and non-functional requirements from clinical and technical stakeholders; (2) design and development of the system architecture using Python, OpenCV, MediaPipe, NumPy, and Tkinter; and (3) clinical usability evaluation in a non-probabilistic sample of 14 patients selected from 30 anophthalmic patients treated at the Cartagena Ophthalmology Clinic between January and July 2025. Eligible participants had unilateral ocular loss, were scheduled for ocular prosthesis fitting, had not previously used an ocular prosthesis, and provided written informed consent. The assessment instrument collected sociodemographic data, previous knowledge, perceived insecurities, support needs, and user evaluation of the simulator. The software enabled real-time facial detection, ocular landmark localization, laterality selection, mirroring and adaptation of the healthy eye, and visualization of the simulated rehabilitation outcome. Most participants rated the simulation as very realistic (85.7%), reported feeling much more confident after using it (71.4%), and stated that they would definitely recommend the software to other patients (92.9%). These findings suggest that the tool is feasible as a patient-centered support resource for pre-prosthetic counseling, expectation management, and clinical communication, while further studies with larger samples and objective technical metrics are required.