Stabilization Algorithm Based on Improved Motion Model for Jittery Video in Minimally Invasive Surgery

The particularity and complexity of the minimally invasive surgical environment may cause the transmitted surgical video to jitter. For jittery video in minimally invasive surgery (MIS), this paper proposed a video stabilization method based on feature matching and improved affine motion model. Firstly, the feature points are extracted and matched using Speeded Up Robust Features (SURF) algorithm. Secondly, for the cumulative error problem, an improved motion model is established on the basis of sequential reference frame which extends adjacent two frames to adjacent multiple frames. Lastly, adjacent frames are transformed and motion compensation is implemented. The experimental results of jittery video in MIS show that the proposed algorithm has high matching precision and small cumulative error.