Simulating Dual-Pixel Images From Ray Tracing For Depth Estimation
This addresses a domain gap problem for researchers and practitioners in computer vision who need accurate depth estimation from DP sensors, but it is incremental as it builds on existing simulation approaches.
The paper tackled the scarcity of dual-pixel (DP) image-depth paired datasets for depth estimation by proposing Sdirt, a simulation method using ray tracing to generate realistic DP images, which improved model generalization to real DP data.
Many studies utilize dual-pixel (DP) sensor phase characteristics for various applications, such as depth estimation and deblurring. However, since the DP image features are entirely determined by the camera hardware, DP-depth paired datasets are very scarce, especially when performing depth estimation on customized cameras. To overcome this, studies simulate DP images using ideal optical system models. However, these simulations often violate real optical propagation laws,leading to poor generalization to real DP data. To address this, we investigate the domain gap between simulated and real DP data, and propose solutions using the Simulating DP images from ray tracing (Sdirt) scheme. The Sdirt generates realistic DP images via ray tracing and integrates them into the depth estimation training pipeline. Experimental results show that models trained with Sdirt-simulated images generalize better to real DP data.