Bayesian classification based pose contour extraction method for moving images

Aiming at the problems of large error and long time-consuming in the traditional motion image attitude contour extraction results, a motion image attitude contour extraction method based on Bayesian classification is proposed. MEMS sensor is used to calibrate the real-time posture of human motion process, determine the posture of moving image, and Bayesian model is used for classification. The log operator is introduced to determine the edge indication function under the adaptive surface evolution extreme value, and combined with the reverse CRF transform to extract the pose contour of the moving image. The experimental results show that the iteration times and time-consuming, signal-to-noise ratio and SSE value of this method are better than those of the traditional method.