We propose an image retrieval framework with the aim of providing visitors of the Museum of Archaeology of the University of Catania (MAUC) with a more immersive and user-friendly experience. Using their smartphone cameras, visitors can take a photo of any artifact and instantly receive detailed information about its shape, place of provenance, ancient function, and more, overcoming the need for descriptive panels or QR codes that may clutter the visual environment. Also, we evaluate several approaches to the image retrieval task, comparing different hand-crafted feature extractors and two descriptors (KDTree and Annoy algorithms). The analysis shows that the Convolutional Neural Network consistently delivers the best performance among the tested techniques. The proposed solution not only improves the accessibility and quality of museum visits but also lays the groundwork for future integration with wearable devices and interactive technologies for cultural heritage exploration.