Novel schemes for high-throughput image rotation

Real-time image rotation forms a core operation in many applications such as medical image processing, computer graphics and pattern recognition. Conventional rotation engines usually comprise of compute-intensive CORDIC operations that become a severe bottleneck in high-throughput applications. In this paper, we present a novel scheme that employs hierarchical computation strategies, exploits the symmetrical characteristics of the image, and performs successive pixel computations at high-speed to achieve high-throughput image rotation at low cost. We show that the proposed method leads to a significant performance gain of about 20 times over an existing CORDIC-based rotation engine with similar hardware cost.