Global Motion Estimation Based on Bit-Plane Matching and Diamond Search
Fang Liu · Ordnance Industry Automation · 2007
A fast global motion estimation algorithm, which combines bit-plane matching method with diamond search strategy, matches sub-images of the frame, and obtains local motion vectors from sub-images. The global motion vector is obtained by selecting medians of each motion vector elements from local motion vectors separately. The matching is realized only by using binary Boolean function. It avoids the affect of unreliable local motion vectors through median filter. Experimental results demonstrate that, compared with full search frame matching algorithm, this algorithm is much faster while it maintains accuracy.