AI-BASED PROPERTY DOCUMENT VERIFICATION SYSTEM

This paper presented an AI-Based Property Document Verification System integrating OCR, structured information extraction, cross-document fuzzy matching, explainable risk scoring and human administrative review. The system brings these steps together in one workflow for detecting inconsistencies in owner names, survey numbers, property areas, PIN codes and supporting documents. In the reported 13-scenario benchmark, all cases were classified according to their expected risk tiers and the measured end-to-end processing time was approximately 1.84 seconds per property. The system demonstrates the feasibility of AI-assisted document analysis for faster property verification while preserving human oversight. Final legal title determination must remain with authorized administrative or judicial processes. Experimental evaluations across 13 rigorous benchmark scenarios demonstrated that the pipeline handles highly complex real-world edge cases—ranging from distorted legacy deeds to sophisticated signature forgeries. Furthermore, the systematic latency performance analysis confirms that the core computational pipeline executes within highly acceptable operating envelopes, striking a functional balance between processing speed and the deep validation checks required for institutional applications

1.  Xu, Y., Li, M., Cui, L., Huang, S., Wei, F., & Zhou, M. (2020). “LayoutLM: Pre-training of Text and Layout for Document Image Understanding.” Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining. https://doi.org/10.1145/3394486.3403172https://www.ecva.net/papers/eccv_2022/papers_ECCV/ html/8042_ECCV_2022_paper.php4.   Jaume, G., Ekenel, H. K., & Thiran, J.-P. (2019). “FUNSD: A Dataset for Form Understanding in Noisy Scanned Documents.” 2019 International Conference on Document Analysis and Recognition Workshops (ICDARW).https://doi.org/10.1109/ICDARW.2019.10029 5.    Mack, C. (2014). “PS1-13: Probabilistic Linkage (Also Known as ‘Fuzzy Matching’): The Theoretical Foundations of Modern Record Linkage.” eGEMs, 2(1). https://doi.org/10.13063/2327-9214.11046.  Wagner, R. A., & Fischer, M. J. (1974). “The String-to-String Correction Problem.” Journal of the ACM, 21(1), 168–173. https://doi.org/10.1145/321796.3218117.  Dwivedi, S. K., et al. (2022). “Blockchain and IPFS-based Reliable Land Registry System.” Security and Privacy. https://doi.org/10.1002/spy2.2368.  Levenshtein, V. I. (1966). “Binary Codes Capable of Correcting Deletions, Insertions, and Reversals.” Soviet Physics Doklady, 10(8), 707–710. Paper details9.   Xu, Y., et al. (2024). “LayoutLLM: Large Language Model Instruction Tuning for Visually Rich Document Understanding.” Proceedings of LREC-COLING 2024. https://aclanthology.org/2024.lrec-main.892/