High Quality Monocular Depth Estimation via Transfer Learning
By Ibraheem Alhashim, Peter Wonka. 2018
499 citations recorded by the source corpus.
Accurate depth estimation from images is a fundamental task in many applications including scene understanding and reconstruction. Existing solutions for depth estimation often produce blurry approximations of low resolution. This paper presents a convolutional neural network for computing a high-resolution depth map given a single RGB image with the help of transfer learning. Following a standard encoder-decoder architecture, we leverage features extracted using high performing pre-trained networks when initializing our encoder along with augmentation and training strategies that lead to more accurate results. We show how, even for a very simple decoder, our method is able to achieve detailed high-resolution depth maps. Our network, with fewer parameters and training iterations, outperforms state-of-the-art on two datasets and also produces qualitatively better results that capture object boundaries more faithfully. Code and corresponding pre-trained weights are made publicly available.
Subfields: Computer Vision and Pattern Recognition, Media Technology, Representation & Self-Supervised Learning
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Computer Vision and Pattern Recognition
- Luc Van Gool
- Jitendra Malik
- Dacheng Tao
- Trevor Darrell
- Alan L. Yuille
- Xiaogang Wang
- Ming-Hsuan Yang
- Andrew Zisserman
In Media Technology
- Alexander Kolesnikov
- Junjun Jiang
- Matthew R. Walter
- André Susano Pinto
- Mohammadreza Mostajabi
- Nicholas I. Kolkin
- Nathan Paczan
- Zhenyu Li
In Representation & Self-Supervised Learning
- Neil Houlsby
- Xiaohua Zhai
- Lucas Beyer
- Alexander Kolesnikov
- André Susano Pinto
- Junjun Jiang
- Nicholas I. Kolkin
- Nathan Paczan
Citations in the corpus
- Hypersim: A Photorealistic Synthetic Dataset for Holistic Indoor Scene Understanding (2020, background): “…the color image. alistic synthetic datasets and interactive simulation environments that have spurred rapid progress towards the goal of holistic indoor scene understanding [6, 7, 9, 10, 14, 15, 19, 22, 24, 31, 33, 36, 37, 39, 43, 44, 45, 46, 49, 52, 61, 62, 63, 65, 70, 72, 75, 76, 79, 83, 84].”
- Clear Grasp: 3D Shape Estimation of Transparent Objects for Manipulation (2019, unknown): “DenseDepth uses a deep neural network to directly predict the depth value from the color image.”
- UViM: A Unified Modeling Approach for Vision with Learned Guiding Codes (2022, methodology): “To contextualize this result, this score is comparable to DenseDepth [1] which uses an architecture composed of a pre-trained DenseNet-169 followed by upsampling layers and skip-connections.”
- DIODE: A Dense Indoor and Outdoor DEpth Dataset (2019, methodology): “Other settings are identical to the original DenseDepth model [3].”
- BinsFormer: Revisiting Adaptive Bins for Monocular Depth Estimation (2022, background): “2, any depth estimation decoder can be adopted ( e.g., [14], [26], [63]).”
- Feedback Adversarial Learning: Spatial Feedback for Improving Generative Adversarial Networks (2019, unknown): “RGB image [2].”