Image Large Rotation and Scale Estimation Using the Gabor Filter

This paper proposes a novel image large rotation and scale estimation method based on the Gabor filter pulse-coupled neural network (PCNN). First, features of template its rotated one are extracted by performing filter. Second, we present modified PCNN model to measure similarity between one. Finally, angle is calculated searching global minimum correlation coefficients. Besides estimation, also propose max-projection strategy. The feature projected along estimated angle, peak this projection result. Moreover, experiments illustrate that proposed has high accuracy robust noise. Compared with state-of-the-art methods, approach more stable performance.