To improve the interframe prediction capability of motion estimation and motion complement (ME+MC) in video coding

a motion pattern recognition (MPR) algorithm based on self-organizing feature maps (SOM) was proposed

and it was tested in object-based conference video coding.The frequency sensitive self-organizing feature maps (FSOM) algorithm was also proposed to overcome the disadvantages of SOM algorithm.Experimental results show that FSOM-MPR algorithm has better coding performances than ME+MC algorithm.When the compression ratio is 170:1

the improvement of PSNR is about 2.7dB using Claire conference video sequence.