We are pleased to announce that Lesław Pawlaczyk (CEO) has become the corresponding author of a scientific publication on the role of automatic analysis of fundus images in detecting lifestyle diseases. The article was presented at the 10th National Scientific Conference "Dimensions of lifestyle and social diseases of the 21st century" on April 11, 2025 in Lublin in online form. The article is titled: "Detection of lifestyle diseases using artificial intelligence based on fundus images", and its abstract is:
The paper presents the process of developing deep neural network models, with particular emphasis on the YOLOv8 architecture, for the analysis of fundus images. The aim was to create a system enabling automatic classification of selected eye diseases: glaucoma, diabetic retinopathy, cataracts and age-related macular degeneration (AMD), as well as prediction of patients' age and gender. The ODIR-5K dataset was used for the experiments, preprocessing, augmentation and division into training, validation and test sets. Three models were built corresponding to the tasks of classification of diseases, age groups and gender, using transfer learning and hyperparameter optimization. The obtained results confirmed the high effectiveness of YOLOv8 in detecting ophthalmic diseases and moderate accuracy in predictive tasks regarding age and gender. The conducted studies emphasize the potential of using artificial intelligence in ophthalmic diagnostics and draw attention to the need for further work on data quality and improving models for telemedicine applications.
This is in line with our broader strategy of publishing scientific papers in the field of ophthalmology and AI. In the coming months, we will inform about further articles.
