Studies On The Database Application In Ship-Manufacture CAM/CAD System & Facial Expression Recognition Technology

In Part 1 of this dissertation, we discuss ship lofting and database technology, as well as their application. It is necessary using database technology to build a ship lofting system. With the knowledge of mathematical fairing and fairing procedure, we divide and classify ship data first; then we design the database, give the tables' structure; after the analyzing and designing of the system, we develop "Ship Lofting CAD/CAM System" with Visual C++ using ODBC technology. The interface of the system is friendly to users. This system can perform fairing automatically and quickly, so it can shorten the process of ship-manufacture remarkably.In part 2 of this dissertation, we briefly discuss the background and the importance of automatic facial expression recognition and generation first. The detail description of this paper is as followed. (1) The recognition of facial expression. We introduce the preprocessing of the facial image including normalization and filtering. Normalization means to enhance the contrast by well-distributed gray degree. Then we introduce edge detection, we combine two operators to get better result. Flexible modules is commonly used in the extraction of the facial organ's character, we simplify it. We roughly locate the organs using horizontal projection and vertical projection of the image, then we locate them exactly using a fast module-matching algorithm. (2) The generation of facial expression. After introducing the characters of the facial expression, the head frame including facial modeling is discussed in detail first, then we construct the parameterization of the head frame, as well as the transform, shading, lighting of the facial model. The frame of the eyes and the mouth are mainly introduced. Based on the theory of the facial action unit, we propose a simplified muscle movement model, in which the movement of the control points reflects the transform of the action units. Finally, the texture mapping method is presented as well as the experiment results.