Bridging the Gap Between AI and Reality • Rhodes, Greece
Time: Wednesday, 5.11
Room: Room C
Authors: Omayma El Haji, Tiziana Margaria
Abstract: The increasing demand for automated document analysis and intelligent image-based applications has highlighted the importance of effective image preprocessing in achieving accurate results in Opti- cal Character Recognition (OCR) systems. However, implementing pre- processing techniques often requires substantial programming expertise, which limitsaccessibilityfornon-technicalusers.Toaddressthis gap,our work explores the integration of image preprocessing steps into a Low- Code/No-Code (LCNC) platform. Specifically, we embed a convolutional neuralnetwork(CNN)-basedOCRmodelwithintheLCNCenvironment, while enabling users to apply and customize essential preprocessing op- erations without the need for extensive coding. The proposed integration provides a streamlined interface that bridges advanced computer vision techniques with user-friendly design. Users can select from a range of preprocessing steps—including noise reduc- tion, normalization, binarization, and contrast enhancement—through intuitive workflows. By combining these preprocessing pipelines with the embedded OCR model, the platform supports flexible adaptation to di- verse document types and real-world conditions. Our results demonstrate that this integration not only improves the usability of OCR and preprocessing functions but also empowers non- programmerstodesign,adapt,anddeploytheirownsolutionsfordomain- specific challenges. Furthermore, by lowering the entry barrier to image processing and OCR, this approach fosters broader adoption of AI tech- nologiesinsectorssuchashealthcare,finance,andeducation.Ultimately, the proposed framework illustrates the potential of LCNC platforms to democratize access to machine learning capabilities, making them more practical, adaptable, and inclusive.