Engineering drawing database retrieval using statistical pattern spotting techniques

Abstract

An experimental mechanical engineering drawing database system, which allows a user to retrieve images by presenting sketches or shapes which represent details such as e.g. screws or holes, is presented in this paper. Due to the use of novel augmented pseudo 2-D HiddenMarkov Models with filler states, images can be retrieved, where the detail that corresponds to the query is embedded in e.g. hatching or is connected to other elements in the image. The proposed technique achieves a good performance which is demonstrated by a number of query and retrieval examples in this paper.