IHIRD: A Data Set for Indian Heritage Image Retrieval

Abstract

Computational approaches are extensively applied to preserve heritage artifacts. Archival of heritage assets is one of its primary focuses, accounting for the ease of accessibility of digital information. To develop an efficient and reliable archival-and-retrieval system of heritage images, it is necessary to have an extensive data set of heritage artifacts, which consists of various kinds of monuments in their digital representation. In this chapter, we develop an image data set, specifically on Indian heritage monuments, called Indian Heritage Image Retrieval Data set (IHIRD), and test it on several retrieval methods. Images of various heritage monuments, like sculptures and paintings, are the elements of this data set. We experimentally evaluate various content-based image retrieval (CBIR) and semantically driven CBIR schemes using this data set and report their performances.