Cultural Heritage; Artificial Intelligence; Handwritten Text Recognition; Image Retrieval
en
Abstract:
The contents of cultural heritage represent a largely untapped treasure, which in the analog age was
accessible only to a limited group of people or existed within their awareness. However, due to the
ongoing digital transformation, numerous barriers are being dismantled in terms of accessibility,
evaluation, and utilization. Through the utilization of cutting-edge AI-based technologies, it
becomes feasible to meticulously visualize these contents, extract and analyze embedded
information, and ultimately interconnect them as desired. Previously relevant factors such as foreign
languages and handwritings are now rendered insignificant, as technology can proficiently recognize
and translate these elements (e.g., translation, HTR, image recognition), thereby converting them
into comprehensible data – resulting in the Big Data of the past.
Standard AI methodologies offer several avenues for enrichment and exploration, including
Handwritten Text Recognition, Symbol Recognition, Image Similarity Search, generating captions
from images (utilizing CLIP – Contrastive Language-Image Pre-Training), and applying ChatGPT to
textual content descriptions.
The objective of this discourse is to delve into the application and potential of AI within the realm of
cultural heritage systems.
en
Research Areas:
Visual Computing and Human-Centered Technology: 100%