Estimating Digital Clay Texture of Mesopotamian Models From 2D Images Using Deep Learning to Render Full- Immersive Virtual Reality (VR)

Archaeologists have introduced AI-powered digital tools to assist in geological surveying of artifacts and identifying their compositional textures, whether clay or rocks, as realistic examples of their ancient settlements. Modern digital applications of virtual and augmented reality are concerned with displaying archaeological models and giving the audience a full immersion that simulates the basic materials from which they were built or carved, and diagnosing the rock or clay components of the earth that were used in their manufacture. Deep convolutional neural network algorithms have played an important role in expanding the capabilities of virtual panoramas by making them more realistic and immersive. In this paper, we produce Mesopotamian deep Panoramic–Virtual Reality (DMP-VR) model for reconstructing a completely immersive, digital clay texture of archaeological models’ information-rich and low-noise super-resolution panoramic scene of the Mesopotamian civilization in 360◦ from low-resolution 2D images of Assyrian and Babylonian models gathered from online search engines. Alignment sensor software compensates for tilt issues during acquisition, and images are first rotated using a geometric transformation depending on the data center image stitching involving cutting training images to random f_sub×f_sub-pixel sub-images and stitching the ends of features to minimize errors. The quantitative and visual comparison of our method (DMP-VR) with other methods achieved ideal results in terms of reconstructing a super-resolution, fully immersive 360-degree panoramic scene. The artifacts have a digital clay texture that is identical to their original reality.