Toward Better SSIM Loss for Unsupervised Monocular Depth Estimation
This work addresses a specific issue in unsupervised depth estimation for computer vision applications, but it is incremental as it modifies an existing loss function rather than introducing a new paradigm.
The paper tackled the problem of improving unsupervised monocular depth estimation by proposing a new form of the SSIM loss function, which uses addition instead of multiplication to combine components, resulting in smoother gradients and achieving higher performance, such as outperforming the baseline on the KITTI-2015 dataset.
Unsupervised monocular depth learning generally relies on the photometric relation among temporally adjacent images. Most of previous works use both mean absolute error (MAE) and structure similarity index measure (SSIM) with conventional form as training loss. However, they ignore the effect of different components in the SSIM function and the corresponding hyperparameters on the training. To address these issues, this work proposes a new form of SSIM. Compared with original SSIM function, the proposed new form uses addition rather than multiplication to combine the luminance, contrast, and structural similarity related components in SSIM. The loss function constructed with this scheme helps result in smoother gradients and achieve higher performance on unsupervised depth estimation. We conduct extensive experiments to determine the relatively optimal combination of parameters for our new SSIM. Based on the popular MonoDepth approach, the optimized SSIM loss function can remarkably outperform the baseline on the KITTI-2015 outdoor dataset.