Art Detective is an AI-powered museum app designed to make art exploration interactive, educational, and playful.
By implementing image recognitionthe app lets users scan artworks, discover details and information about the painting, the artist, and experience museums in a more immersive way.
GitHub repository
We began the project by studying existing art and museum applications to identify gaps and opportunities. The goal was to find ways to combine image recognition, storytelling, and generative design while keeping the experience engaging for museum visitors.
The analysis revealed opportunities to improve the Rijksmuseum app through faster image recognition, personalized recommendations, and interactive features such as filters and generative art expansions. These insights shaped our approach to make the Art Detective concept both informative and entertaining.
We defined the project through “How Might We” (HMW) questions that guided both the design and technical aspects of the app. Each question aimed to balance creativity, functionality, and privacy within the museum context.
We explored ways to make the app stand out through playful interaction, such as integrating mini-games, fun facts, or animations that appear when scanning artworks. For image recognition, we evaluated several tools and selected Google Vision API for its accuracy and integration flexibility with the Rijksmuseum API. Finally, we addressed privacy concerns by processing scans in real time without cloud storage or saving user data.
Before designing, we defined the logical structure of the app — starting from the welcome screen to the artwork scan and information display. The goal was to create a simple and intuitive flow that supports both casual exploration and educational discovery.
The app’s flow focused on three key moments: onboarding, scanning, and learning. Visitors first log in or access as guests, then scan an artwork using the camera. The app processes the image, retrieves information, and presents it with dynamic visual layers.
Through multiple design iterations, we refined both the layout and the information structure. Each round improved usability, clarity, and overall aesthetic consistency.
In the first version, we had a simple welcome screen and scanning interface. The second iteration introduced a dedicated login and a more informative home view, while the third version refined the layout with rounded artist icons, restructured data panels, and simplified the navigation to focus on active engagement during scans.
How our design might help Rijksmuseum?
We integrated Google Vision API for artwork recognition. The system identifies artworks in real time and cross-references results with the Rijksmuseum API to deliver contextual information and related works instantly.
As the part of our project we created a funny promotional video showcases the Art Detective concept.