Learning Depth from Single Monocular Images

By Ashutosh Saxena, Sung H. Chung, Andrew Y. Ng. 2005

1,230 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In 3D Vision & Pose

Citations in the corpus

Hierarchical Surface Prediction for 3D Object Reconstruction (2017, background): “5D setting where the goal is, given a color image to predict a depth map [11], [21], [30].”
3D Shape Induction from 2D Views of Multiple Objects (2016, methodology): “Recognition-based methods have been used to estimate geometry of out-door scenes [14, 26], indoor environments [9, 27], and objects [1, 25].”
Depth Map Prediction from a Single Image using a Multi-Scale Deep Network (2014, methodology): “[15] predict depth from a set of image features using linear regression and a MRF, and later extend their work into the Make3D [16] system for 3D model generation.”
(CAD)$^2$RL: Real Single-Image Flight without a Single Real Image (2016, background): “in the hallways, more detailed depth estimation [36, 24, 23] could also be used without loss of generality.”
Deep convolutional neural fields for depth estimation from a single image (2014, methodology): “Furthermore, these methods suffer from lacking of flexibility in that [14,15] rely on horizontal alignment of images and [7] requires the semantic labellings of the training data available beforehand.”
Deeper Depth Prediction with Fully Convolutional Residual Networks (2016, methodology): “Classic methods on monocular depth estimation have mainly relied on hand-crafted features and used probabilistic graphical models to tackle the problem [8, 17, 29, 30], usually making strong assumptions about scene geometry.”
Pixel Recursive Super Resolution (2017, background): “super resolution [31], colorization [50, 6], and depth estimation [37], models that are able to predict a single mode are heavily preferred over models that blend modes together.”
Fast End-to-End Trainable Guided Filter (2018, methodology): “In Computer Vision, the pixelwise dense prediction techniques not only address the problem of segmenting an image into semantic parts [22, 35, 11], but also help to estimate depth from a single image [34], and detect the most salient object in an image [32, 29].”
FastDepth: Fast Monocular Depth Estimation on Embedded Systems (2019, methodology): “crafted feature representations [6] to modern deep learning based approaches [11–14].”
BA-Net: Dense Bundle Adjustment Network (2018, methodology): “Before the raise of deep learning based methods, some works predict depth from a single image based on MRF (Saxena et al., 2005; 2009), semantic segmentation (Ladický et al.”
Deep Depth Completion of a Single RGB-D Image (2018, methodology): “Other early methods were based on hand-tuned models and/or assumptions about surface orientations [31, 60, 61].”
Single Image Dehazing via Conditional Generative Adversarial Network (2018, methodology): “In addition, the Make3D datasets [27, 28, 29] are employed as the outdoor images.”
Depth-Aware Video Frame Interpolation (2019, background): “Recently, several learning-based approaches [8, 9, 11, 18, 20, 31, 32, 37] aim to estimate the depth from a single image.”
Neural Window Fully-connected CRFs for Monocular Depth Estimation (2022, methodology): “[30] claim that local features alone are insufficient to predict the depth of a pixel, and the global context of the whole image needs to be considered to infer the depth.”
FD-GAN: Generative Adversarial Networks with Fusion-discriminator for Single Image Dehazing (2020, background): “However, most of the dehazing data are synthesized from NYU dataset (Silberman et al. 2012), which only includes indoor images.”
Recovering Occlusion Boundaries from an Image (2011, background): “[62, 63] train with range images to estimate depth from the image features directly.”
Depth Transfer: Depth Extraction from Video Using Non-Parametric Sampling (2014, background): “Saxena et al. devised a supervised learning strategy for predicting depth from a single image [7], which was further improved to create realistic reconstructions for general scenes [8], and efﬁcient learning strategies have since been proposed [9].”
Monocular Depth Estimation Using Neural Regression Forest (2016, methodology): “We use three standard error metrics which are commonly used by the state of the art methods [21, 22, 14, 6] to estimate the accuracy of monocular depth prediction.”