Back close

UNet and Dense Extreme Inception Networks: A Unified Approach to Archeological Artifact Drawing Generation

Publication Type : Conference Proceedings

Publisher : IEEE

Source : 2025 International Conference on Computing for Sustainability and Intelligent Future (COMP-SIF)

Url : https://doi.org/10.1109/comp-sif65618.2025.10969917

Campus : Bengaluru

School : School of Engineering

Department : Electronics and Communication

Year : 2025

Abstract : Archaeological artifact drawing is a crucial yet labor-intensive task in cultural heritage preservation and research. This paper presents a unified approach that integrates UNet and Dense Extreme Inception Network (DEIN) architectures to automate the generation of artifact drawings with high precision and efficiency. The proposed framework leverages the feature extraction capabilities of DEIN to capture multi-scale and contextual information, while UNet's encoder-decoder structure ensures accurate localization and detailed reconstruction of artifact outlines. Intermediate feature maps are processed through a series of upsampling networks, enabling the generation of multilevel edge maps, which are subsequently fused to produce a single, refined output. The effectiveness of this approach is demonstrated through large number of experiments on a Corpus of the Stampseals of the Southern Levant (CSSL) dataset of archaeological artifacts. Experimental results demonstrate that the unified architecture outperforms existing methods by effectively preserving fine details and capturing complex structures in artifact drawings. This work provides an automated and scalable solution for digitizing archaeological documentation, reducing manual effort while enhancing the accuracy of artifact representations.

Cite this Research Publication : Sushma B, UNet and Dense Extreme Inception Networks: A Unified Approach to Archeological Artifact Drawing Generation, 2025 International Conference on Computing for Sustainability and Intelligent Future (COMP-SIF), IEEE, 2025, https://doi.org/10.1109/comp-sif65618.2025.10969917

Admissions Apply Now