<![CDATA[Optical Character Recognition (OCR) plays a key role in this process by converting text contained in scanned documents into machine-readable information. However, OCR performance is highly dependent on document image quality, which may be affected by factors such as blur, low illumination, physical deterioration, and perspective distortion. This study evaluates the performance of the OCR module implemented in the SmartScan Rivai application under varying document quality conditions. Five document conditions were considered: normal, blurred, worn, low illumination, and skewed perspective. The evaluation focused on three performance indicators: Customer ID recognition accuracy, document classification accuracy, and OCR processing time. Experimental results show that the OCR system successfully recognized Customer IDs in 22 out of 25 test cases, achieving an overall accuracy of 88%. The highest recognition accuracy (100%) was obtained for normal and worn documents, whereas blurred documents, low illumination, and skewed perspectives reduced the accuracy to 80%. Document classification achieved an overall accuracy of 67%, indicating that this task is more challenging because it depends on the successful recognition of multiple textual features rather than individual characters. In addition, all OCR processes were completed in less than five seconds per document, demonstrating the operational feasibility of the proposed system. The findings confirm that document quality significantly influences OCR performance and highlight the importance of incorporating image preprocessing techniques to improve recognition accuracy under challenging document conditions. Overall, SmartScan Rivai provides an effective solution for operational document digitization while offering opportunities for further enhancement through advanced image processing and artificial intelligence-based OCR techniques.]]>
