Assessing the Effectiveness of Image Hashing and Vector Databases in Improving the Detection of Financial Statement Fraud in the Oil and Petroleum Sector

Abstract Financial statement fraud in the oil and petroleum sector poses a significant challenge, necessitating advanced detection methods. This study evaluates the effectiveness of integrating techniques from Approach 1 and methods from Approach 2 to enhance fraud detection in scanned financial documents. The research addresses limitations of existing rule-based algorithms by adopting a novel hybrid approach. The methodology incorporates an encryption technique to generate unique cryptography for document images, enabling fraud evidence identification through cryptography comparisons. This is complemented by generative AI and deep learning techniques to refine the filtering process and uncover hidden fraud patterns. The training dataset is enriched with various fraud combinations, and the model's performance is assessed using established metrics. Post-processing involves employing image fuzzy logic to mitigate potential inaccuracies. Results indicate that this integrated approach significantly surpasses traditional image processing techniques in both accuracy and detection speed. The deep learning AI component demonstrates particular efficacy in identifying concealed fraud patterns within scanned documents. Furthermore, the model's ability for continuous learning enables adaptation to emerging fraud patterns through periodic retraining with updated document sets. The study's novelty lies in its systematic evaluation and comparison of combined traditional and advanced AI techniques for fraud detection in the oil and petroleum industry, addressing a critical research gap. The implementation of an AI-driven learning mechanism for ongoing refinement further enhances the model's maturity and accuracy over time, contributing to the establishment of a robust and adaptive fraud detection system. This research constitutes a notable contribution to the field of financial fraud detection, presenting a promising avenue for enhancing security and integrity in financial reporting within the oil and petroleum sector.