Hello, my research is about sign language recognition, many researchers choose to use the sign as a base unit of modeling , while others attempt to use a structure similar to phonemes to create models. what's the better approach for modeling the sign?
This depends on the signs that you have in a particular sign language. if a sign language has a large set of signs that the end position of a hand movement define a sign, then entire sign can be taken as one entity and try to recognize it.
otherwise, determination of sign as set of of different hand/finger positions will be the way to go. most probably you may have to use either statistical or AI based techniques to build a hypothesis on whether a bone is within the range required by a particular sign or not because of the variations of signing of a single sign by two different people.
Hope this helps .. I am engaged in a research to do the exact opposite of what you do that is animating signs for Sinhala sign language :D
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Human-Computer Interaction