Toward Better SSIM Loss for Unsupervised Monocular Depth Estimation

2025-06-05

International Conference on Image and Graphics

AI Summary

Conventional SSIM loss in unsupervised monocular depth estimation suffers from gradient instability and poor robustness due to multiplicative coupling among luminance, contrast, and structure components. Method: We propose a novel additive SSIM loss that replaces the multiplicative combination with an additive one—first of its kind—and systematically optimize component weights to better model challenging regions such as textureless areas and motion boundaries. Integrated into the MonoDepth framework, our method is trained end-to-end with photometric consistency constraints, and hyperparameter design is guided by parameter sensitivity analysis. Results: On KITTI-2015, our approach significantly outperforms the baseline, reducing absolute relative error (AbsRel) by 8.2%. Notably, improvements are most pronounced in low-texture and dynamic-edge regions, demonstrating enhanced generalization and robustness under challenging conditions.

Technology Category

• Computer Vision: Learning & Optimization for CV
• Search and Optimization: Learning to Search
• Machine Learning: Multimodal Learning

Application Category

• Responsible Web: Human-perceived consequences of algorithmic deployment on the web
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

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.

Problem

• Improving SSIM loss for unsupervised monocular depth estimation
• Addressing component and hyperparameter effects in SSIM function
• Enhancing depth estimation performance with optimized SSIM loss

Innovation

• New SSIM form with additive component combination
• Smoother gradients for better depth estimation
• Optimized parameters for higher performance