Interactive OCR Editor for Image-Based Text Editing
Nishitha Kanakanapuri, Likhitha Meher Medisetty, Sudeeshna Bhutham, Premkumar Chithaluru · International Journal of Research Publication and Reviews · 2025
Text editing on images is usually difficult to achieve since it involves difficult dynamic interactions with characters and digits which are usually embedded.The project herein proposes an interactive system for text editing based on the Optical Character Recognition (OCR) method to overcome such a problem.The system performs preprocessing of input images via grayscale conversion and adaptive thresholding to increase the visibility of text.The Tesseract OCR engine is used for fetching textual content along with bounding box coordinates for characters and digits.A novel interactive editing interface is proposed where users can modify the detected text directly on the image.In this system, a bounding box-based method of selection is used for locating the text regions.When the user interacts, the updated text is seamlessly reinserted into the image.The utilization of Tkinter for real-time rendering facilitates seamless and intuitive editing capabilities.The suggested methodology has undergone assessment across a range of input images to validate its effectiveness and precision.Metrics such as the accuracy of character detection, latency of edit updates, and responsiveness of the user inter-face serve to highlight the system's resilience.The initiative presents an effective and accessible method for changing embedded text in images, thereby applicable to correction of scanned documents, digitized forms, and improving annotated images.