Abstract: | This letter presents a novel approach for organizing computational resources into groups within H.264/AVC motion estimation architectures, leading to reductions of up to 75% in the equivalent gate count with respect to state‐of‐the‐art designs. |