DepthCues: Evaluating Monocular Depth Perception in Large Vision Models
Authors:
Preprint
Abstract
Large-scale pre-trained vision models are becoming increasingly prevalent, offering expressive and generalizable visual representations that benefit various downstream tasks. Recent studies on the emergent properties of these models have revealed their high-level geometric understanding, in particular in the context of depth perception. However, it remains unclear how depth perception arises in these models without explicit depth supervision provided during pre-training. To investigate this, we examine whether the monocular depth cues, similar to those used by the human visual system, emerge in these models. We introduce a new benchmark, DepthCues, designed to evaluate depth cue understanding, and present findings across 20 diverse and representative pre-trained vision models. Our analysis shows that human-like depth cues emerge in more recent larger models. We also explore enhancing depth perception in large vision models by fine-tuning on DepthCues, and find that even without dense depth supervision, this improves depth estimation. To support further research, our benchmark and evaluation code will be made publicly available for studying depth perception in vision models.Keywords
All Related Versions
This publication has 41 references indexed in Scilit:
- On the Viability of Monocular Depth Pre-training for Semantic Segmentation2024
- CONDENSE: Consistent 2D/3D Pre-training for Dense and Sparse Features from Multi-View Images2024
- Improving 2D Feature Representations by 3D-Aware Fine-Tuning2024
- MoVideo: Motion-Aware Video Generation with Diffusion Model2024
- DeiT III: Revenge of the ViT2022
- Semantic Amodal Segmentation2017
- Unsupervised CNN for Single View Depth Estimation: Geometry to the Rescue2016
- Deep Residual Learning for Image Recognition2016
- Describing Textures in the Wild2014
- Are we ready for autonomous driving? The KITTI vision benchmark suite2012