http://dx.doi.org/10.1016/j.cviu.2010.03.004 We propose confocal conics as a new planar pattern for camera calibration with poses. We show that confocal conics encode the metric information which can be easily extracted with their generalized eigenvalues. The proposed method is tested with synthetic and real data, and shows comparable performances with conventional algorithms, while it still has all virtues of curved features.