DocMaster is a new system designed to analyze complex documents like academic papers and financial reports by preserving their original hierarchical structures, such as sections, tables, and figures. Unlike existing methods that flatten documents into plain text, DocMaster parses them into hierarchical trees and builds a structure-aware semantic index to improve filtering and analysis accuracy.
- Parses documents into hierarchical trees while preserving original layouts.
- Constructs a structure-aware semantic index for accurate document filtering.
- Enables in-depth question answering over filtered results via natural language.
- Provides an interactive web interface for uploading collections and building indices.
The system addresses the performance degradation caused by discarding rich structural information, offering users a way to conduct more precise analysis on selected document subsets.