A Line Drawings Degradation Model for Performance Characterization

Line detection algorithms constitute the basis for technical document analysis and recognition. The performance of these algorithms decreases as the quality of the documents degrades. To test the robustness of line detection algorithms under noisy circumstance, we propose a document degradation mode, which simulates noise types that drawings may undergo during their production, storage, photocopying, or scanning. Using our model, a series of document images at various noise levels and types can be generated for testing the performance of line detection algorithms. To illustrate that our model is consistent with real world noise types, we validated the method by applying it to three line recognition algorithms.