This paper presents a novel method for recognizing expiration dates that is optimized for the structural complexity and inference speed on food packages, designed to assist vision-impaired individuals. Utilizing computer vision techniques, specifically convolutional neural networks (CNNs), the method efficiently extracts date components (day, month, and year) and converts them into a readable date-time format. Validated on real-world datasets, the proposed method shows significant speed improvements, achieving a precision of 0.9850, a recall of 0.8814, and an F1 score of 0.9303, which is equivalent to the state of the art. The approach demonstrates superior performance compared to existing expiration date recognition systems, ensuring fast inference and autonomy for vision-impaired users.