Aberration-Aware Depth-from-Focus
This addresses a specific problem in computer vision for applications requiring precise depth estimation, but it is incremental as it builds on existing methods.
The paper tackles the domain gap in depth estimation caused by lens aberrations in Depth-from-Focus tasks, showing that aberration-aware training improves accuracy without fine-tuning or architectural changes.
Computer vision methods for depth estimation usually use simple camera models with idealized optics. For modern machine learning approaches, this creates an issue when attempting to train deep networks with simulated data, especially for focus-sensitive tasks like Depth-from-Focus. In this work, we investigate the domain gap caused by off-axis aberrations that will affect the decision of the best-focused frame in a focal stack. We then explore bridging this domain gap through aberration-aware training (AAT). Our approach involves a lightweight network that models lens aberrations at different positions and focus distances, which is then integrated into the conventional network training pipeline. We evaluate the generality of pretrained models on both synthetic and real-world data. Our experimental results demonstrate that the proposed AAT scheme can improve depth estimation accuracy without fine-tuning the model or modifying the network architecture.