A low-power video motion estimation array processor

For remote-sensing and portable video encoding applications, reducing power consumption is important due to the limited power budget. In systems such as MPEG, motion estimation (ME) accounts for most of the compression and computation. In ME processing, each image frame is broken into macro-blocks (MB) for searching in the previous or next frame using mean absolute difference (MAD) as the metric function. When a good match is found, the motion vector and error block are sent. Smart searches have fewer MAD operations than brute-force search, saving power with some loss in compression ratio. The brute-force and a variation of 2Log algorithms are compared in this paper.

[1]  Teresa H. Meng,et al.  Parallel array architectures for motion estimation , 1991, Proceedings of the International Conference on Application Specific Array Processors.

[2]  Masahiko Yoshimoto,et al.  A half-pel precision motion estimation processor for NTSC-resolution video , 1993, Proceedings of IEEE Custom Integrated Circuits Conference - CICC '93.

[3]  F. Jutand,et al.  A versatile and powerful chip for real-time motion estimation , 1989, International Conference on Acoustics, Speech, and Signal Processing,.

[4]  J. Shott,et al.  A 200 mV self-testing encoder/decoder using Stanford ultra-low-power CMOS , 1994, Proceedings of IEEE International Solid-State Circuits Conference - ISSCC '94.