Expert system for the visual recognition of arm movements
Embargo Date
Indefinite
OA Version
Citation
Abstract
Based on experience, humans have the ability to anticipate the continuation of a movement after they have seen and recognized part of it. An expectation-driven visual recognition system similarly deals with the problems of having a computer make perceptual predictions of motion. In this work, an expectation-driven expert system for the visual recognition of arm movements is proposed. The system relies on the variations of the joint angles to make predictions and recognize motion. It takes as input the waveforms describing the variations of those angles, segments them at multiple levels and generates symbolic movement descriptions. As the motion continues, it updates the knowledge-base with new motion descriptors and makes use of the information already acquired or inferred to proceed. It uses the initial motion data to classify the movement at hand and the following information to identify it.
Description
Dissertation (Ph.D.)--Boston University
License
PLEASE NOTE: Boston University Libraries did not receive an Authorization To Manage form for this thesis or dissertation. It is therefore not openly accessible, though it may be available by request. If you are the author or principal advisor of this work and would like to request open access for it, please contact us at open-help@bu.edu. Thank you.