The repository contains the source code for a cross-platform mobile app (built using Flutter) for the Cypriot National Gallery of Modern Art in Nicosia, Cyprus.
The app was built by the BIO-SCENT research group at CYENS Centre of Excellence, as part of the published study Theodosiou et al. (2022) "A Systematic Approach for Developing a Robust Artwork Recognition Framework Using Smartphone Cameras".
The app is able to automatically identify a number of artworks at the Gallery through the use of Computer Vision, and display details about them to the user. The artwork identification is performed live on-device, using a Convolutional Neural Network (CNN) that trained using transfer learning on MobileNet V2.
The repository is split into 2 parts:
- The source code for the app demo, built using Flutter, and
- The source code used to train and validate the neural networks, in Python.
Pre-built .apk files for the app can be downloaded and run on any Android device from the Releases section on the sidebar.
