#209
Depth Anything is a highly practical solution for robust monocular depth estimation. It is a simple yet powerful foundation model dealing with any images under any circumstances. The authors scale up the dataset by designing a data engine to collect and automatically annotate large-scale unlabeled data (~62M), which significantly enlarges the data coverage and thus is able to reduce the generalization error. Two simple yet effective strategies that make data scaling-up promising. First, a more challenging optimization target is created by leveraging data augmentation tools. It compels the model to actively seek extra visual knowledge and acquire robust representations. Second, an auxiliary supervision is developed to enforce the model to inherit rich semantic priors from pre-trained encoders. Its zero-shot capabilities are evaluated on six public datasets and randomly captured photos. It demonstrates impressive generalization ability. Further, through fine-tuning it with metric depth information from NYUv2 and KITTI, new SOTAs are set. This better depth model also results in a better depth-conditioned ControlNet.
In this video, I talk about the following: What is monocular depth estimation? How is the Depth Anything model trained? How does Depth Anything perform?
For more details, please look at https://arxiv.org/pdf/2401.10891 and https://github.com/LiheYoung/Depth-An...
Yang, Lihe, Bingyi Kang, Zilong Huang, Xiaogang Xu, Jiashi Feng, and Hengshuang Zhao. "Depth anything: Unleashing the power of large-scale unlabeled data." In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 10371-10381. 2024.