Real-time endpoint detection of upper limb movement based on energy threshold and residual methods

Collection of upper-limb movement acceleration information, handled by smoothing filtration, is used to find the endpoint in the real-time detection. We extracted a part of the collection by a new technology combining the energy-threshold detection technology and the residual error technology. This part contains single complete motion information. Then, this part extracted will be handled by normalization and interpolation for action recognition. Compared with the previous method, this method can be effective for real-time action recognition.