IRIS

The IRIS project aims to enhance the relevance and effectiveness of digital signage in (semi-) public spaces. The goal is to create and demonstrate an adaptive and engaging communication experience that adapts to customers’ emotional and cognitive states in real-time, all while maintaining privacy compliance.

The IRIS partners plan to achieve this by leveraging neuro-scientific insights and advanced image processing techniques.

Combining physiological tracking with AI

At the heart of the IRIS project are non-intrusive visual physiological tracking methods, such as gaze analysis, pupil dilation, facial expressions, and body movement, which provide clues about the viewer’s state of mind.

By integrating these tracking data with generative AI, the project aims to tailor dynamic content to each individual or group in a privacy-preserving manner.

The project will deliver a real-life demonstrator showcasing the future of personalized, context-aware digital communication in public and semi-public spaces, while also demonstrating its commercial viability.

Given the project's reliance on personal data, privacy and ethical considerations are paramount. The IRIS consortium is committed to developing a framework that ensures data minimization and privacy by design, addressing the privacy-relevance paradox. The project will explore visual consent capture mechanisms, data minimization strategies, and compliant data processing methods that adhere to GDPR.

Research goals The project revolves around five key research goals:

1. Physiological Feature Extraction: Using deep learning models to track eye movements, facial features, and body gestures, the project seeks to design a model capable of detecting gaze, pupil size, and other physiological responses from regular RGB cameras.
2. User Segmentation and Machine Learning: The extracted physiological data will be used to segment users based on their emotional and cognitive states, creating different consumer profiles.
3. Content Adaptation: The project will develop algorithms that translate neuroscientific data into content distribution strategies. This involves adjusting message type, format, and repetition, based on passers-by's emotional and cognitive states.
4. User Experience and Acceptance: A key component of the project is understanding how users interact with this adaptive technology.
5. Privacy and GDPR Compliance: The project will focus on creating privacy-preserving data processing methodologies that respect legal constraints while maintaining the effectiveness of the personalized content.

“The IRIS project addresses current technological and market gaps and lays the groundwork for future innovations in smart city applications and retail marketing.”