<![CDATA[Business cards serve as a form of identification that facilitates communication, but managing large amounts of contact information on business cards can often be challenging. To address this issue, this study developed an end-to-end architecture model to automatically extract information from business cards image. This model utilizes Optical Character Recognition and the Faster Region-Based Convolutional Neural Network method. This model allows users to extract contact information from business cards. Using a dataset of 450 business card images, we conducted experiments to evaluate their impact on the task of detecting text in images. We used an image batch size of 500 with 50, 100, and 500 epochs as hyper experiment parameters. The highest accuracy achieved was 0.8342 with mAP was 0.8513. For the character recognition task, Optical Character Recognition produced results with a Character Error Rate (CER) less than 0.08. These findings suggest that the integration of Faster R-CNN and OCR is effective in detecting and extracting textual content from diverse business card layouts. In conclusion, the proposed approach provides a reliable and efficient solution for automated business card digitization and shows strong potential for practical applications in contact information management systems.]]>
Copyrights © 2026
