Research on Unstructured Electronic Archives Query Based on Visual Retrieval Technology

Abstract

Digitalization of enterprise archives is the mainstream trend of archive management. This paper proposes a digital archive index management framework based on visual retrieval technology for unstructured digital archive management problems. The framework adopts the current mainstream deep local feature extraction scheme DELF Pipeline to carry out feature extraction for digital archives, and use the distributed inverted indexing framework Lucene to build an efficient indexing and retrieval system for digital archives. Through a large number of simulation experiments, it is proved that the framework can be well used for the management of enterprise unstructured digital archives, which supports dynamic incremental index construction and has high retrieval efficiency.