Research Of The Retrieval Techniques For Motion Data

For computer generated character animation the character model is driven by motion data which generally is acquired through motion synthesis, editing and capture. Since the volume of motion data grows larger motion retrieval technology becomes a critical part of character animation. In this paper we first discuss how the motion is recorded as motion data and its parameterized representation. An intuitive way of measuring motion similarity is to use numerical comparison of motion data and is applied in a hit and reaction motion synthesis system.Motion retrieval could also be carried out in other forms such as movement notation language. We implement a framework which allows the user to retrieve motions via Labanotation. A similarity metric for Labanotation sequences is proposed and used to search the motions that have similar Laban descriptions. Our search algorithm is able to retrieve motion segments that only match part of the query Laban sequence. Then based on dynamic programming these segments are stitched together to form a smooth output motion that is in an optimal sense of matching query Laban sequence. Experimental results demonstrate our method could effectively improve the utilization of motion data.Sketch-drawing is another intuitive and comprehensive means of conveying movement ideas. We proposed a novel sketch-based approach to retrieve desired Kungfu motions. Given two human figure sketches corresponding to the initial and closing posture of a Kungfu form, and the trajectory drawings on specific moving joints, a system named MotionMaster can find prototype motion by sketch-based retrieval in a motion database. It greatly helps the animator focus on the movement idea development during the evolutionary process of building motion data for articulated characters. These three kinds of motion retrieval techniques are discussed in chapter 3,4 and 5 respectively.