Year: 2025
Type: article
Abstract: Dark patterns in user interfaces are subtle yet manipulative design techniques that deceive users into performing unintended actions, compromising their experience and trust. Our project aims to revolutionize user interaction by developing an automated system that detects, analyses, and resolves these dark patterns in mobile and web applications. Leveraging Fast Region-based Convolutional Neural Networks (FRCNN), our system identifies deceptive visual elements, while advanced bounding box detection and EAST text identification techniques extract and analyse UI elements and text content to pinpoint misleading messages. By incorporating text pattern matching, colour brightness analysis, and spatial analysis, we uncover hidden dark patterns and provide users with real-time notifications, fostering transparency and trust. Our approach includes strategies to mitigate and correct these patterns, promoting ethical design practices and transforming the digital landscape into a more trustworthy environment. Through rigorous testing and validation, we ensure the reliability and effectiveness of our system, contributing to a safer, more honest user experience.
Author: Arivumathi V
Language: English
Subfield: Human-Computer Interaction
Field: Computer Science
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W4409501388