<p>Visually impaired individuals face significant challenges in distinguishing Egyptian banknotes due to their similar sizes, colors, and textures. While many studies have examined currency recognition, most focus on widely used global currencies such as USD, EUR, or INR and rely on traditional machine learning or early CNN architectures. Very limited research has investigated Egyptian currency, and existing methods often struggle with issues like worn notes, changing lighting conditions, and complex local designs. Additionally, no previous study has integrated modern object detection frameworks like YOLOv8–YOLOv10 or used metaheuristic optimization to improve model accuracy. To close these research gaps, this study presents a novel Egyptian currency recognition system employing advanced YOLO architectures optimized for accuracy and computational efficiency. A dataset of 2,000 annotated Egyptian banknote images was used to train and assess the models. Innovations such as context aggregation, GELAN, and NMS-free training were implemented, and Harris Hawks Optimization (HHO) was utilized to fine-tune YOLOv8’s learning rate, increasing its validation mAP@0.5 from 0.9707 to 0.9781. Experimental results show that YOLOv10 achieved the best performance, with a precision of 0.9678, F1 score of 0.9715, and mAP@0.5 of 0.9934. The proposed framework addresses the notable gap in Egyptian currency research by combining deep learning advancements and optimization techniques, providing a robust and inclusive AI solution that empowers visually impaired users.</p>