Pioneering Simplicity in Computer Vision: A First-Hand Account of Developing a Low-Code OpenCV Interface for Defect Detection Applications

In the pursuit of democratizing computer vision technology, this paper presents a first-hand account of the development of a low-code interface for OpenCV, a leading library in the field. Our research introduces a user-friendly platform that enables users with varying levels of programming expertise to effectively implement computer vision solutions. Recognizing the technical barriers presented by traditional hard coding, we have integrated a meticulous survey method designed to assess the effectiveness of the low-code interface for defect detection tasks within computer vision. Participants from diverse expertise levels-Beginners, Intermediates, and Experts-were provided with three defect detection use cases. They were tasked with identifying and delineating defects using both traditional OpenCV methods and our innovative low-code GUI and SDK, enabling a direct comparison of the two approaches in a controlled environment. By analyzing the development time and parameters tuning, we demonstrate significant reductions in implementation time across all levels of expertise. For hard code implementations, with Beginners reducing their development time from 39.4 minutes to just 5.9 minutes, Intermediates from 24.5 minutes to 3.4 minutes, and Experts from 9.1 minutes to a brisk 1.3 minutes. Moreover, parameter tuning-often a time-intensive process-also saw considerable time decreases, with Beginners' times dropping from 25.4 to 15.0 minutes, Intermediates from 18.0 to 10.0 minutes, and Experts from 10.2 to 5.3 minutes when using the low-code solution. These results underscore the potential of low-code platforms to significantly streamline the development process, simplifying complex tasks and offering a more inclusive and productive environment for developers of all levels. The findings advocate for a paradigm shift towards low-code platforms, especially in fields that demand rapid and agile development cycles like computer vision.