Network trained to estimate 4 parameters - roll, distorted offset, field of view, k1_hat - on images with varying aspect ratios
Network trained to estimate 4 parameters - roll, distorted offset, field of view, k1_hat - on images with varying aspect ratios