We introduce SEA-RAFT, a more simple, efficient, and accurate RAFT for optical flow. Compared with RAFT, SEA-RAFT is trained with a new loss (mixture of Laplace). It directly regresses an initial flow for faster convergence in iterative refinements and introduces rigid-motion pre-training to improve generalization. SEA-RAFT achieves state-of-the-art accuracy on the Spring benchmark with a 3.69 endpoint-error (EPE) and a 0.36 1-pixel outlier rate (1px), representing 22.9% and 17.8% error reduction from best published results. In addition, SEA-RAFT obtains the best cross-dataset generalization on KITTI and Spring. With its high efficiency, SEA-RAFT operates at least 2.3x faster than existing methods while maintaining competitive performance. The code is publicly available at https://github.com/princeton-vl/SEA-RAFT.

翻译：我们提出了SEA-RAFT，一种更简单、高效且准确的用于光流估计的RAFT模型。与RAFT相比，SEA-RAFT采用了一种新的损失函数（拉普拉斯混合分布）进行训练。它直接回归一个初始光流，以在迭代优化中实现更快的收敛，并引入了刚性运动预训练以提高泛化能力。SEA-RAFT在Spring基准测试上取得了最先进的精度，其端点误差（EPE）为3.69，1像素异常值率（1px）为0.36，相较于已发表的最佳结果，误差分别降低了22.9%和17.8%。此外，SEA-RAFT在KITTI和Spring数据集上获得了最佳的跨数据集泛化性能。凭借其高效率，SEA-RAFT的运行速度至少是现有方法的2.3倍，同时保持了有竞争力的性能。代码公开于 https://github.com/princeton-vl/SEA-RAFT。