Depth Estimation from Image Structure

By Antonio Torralba, Aude Oliva. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2002

385 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition, Media Technology

View Paper: https://www.semanticscholar.org/paper/acf0b6745708457a53d5327eea345c0bcf466e97

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

In Media Technology

Citations in the corpus

Contextual guidance of eye movements and attention in real-world scenes: the role of global features in object search. (2006, background): “For instance, the degree of perspective or the mean depth of the space that a scene image subtends can be estimated by a configuration of low-level image features (Torralba & Oliva 2002, 2003).”
80 Million Tiny Images: A Large Data Set for Nonparametric Object and Scene Recognition (2008, methodology): “Low dimensional global image representation have been used to for scene recognition [16] , [32], [22], for providing context for object detection [38], [40] , depth estimation [ 41 ] and image retrieval for computer graphics [ 20].”
Recognition of natural scenes from global properties: Seeing the forest without representing the trees (2008, neutral): “…without the need for first identifying component objects ([Fei-Fei and Perona, 2005], [Oliva and Torralba, 2001], [Oliva and Torralba, 2002], [Oliva and Torralba, 2006], [Torralba and Oliva, 2002], [Torralba and Oliva, 2003], [Vogel and Schiele, 2007] and [Walker Renninger and Malik, 2004]).”
Putting Objects in Perspective (2006, background): “From low-level image cues, Torralba and Oliva [26] get a sense of the viewpoint and mean scene depth, which provides a useful prior for object detection [27].”
Make3D: Learning 3D Scene Structure from a Single Still Image (2009, background): “Torralba and Oliva [20] studied the relationship between the Fourier spectrum of an image and its mean depth.”
Building the gist of a scene: the role of global image features in recognition. (2006, background): “Building a scene representation from global image features High-level properties of a scene such as the degree of perspective or the mean depth of the space that the scene subtends have been found to be correlated with the configuration of low-level image features (Torralba and Oliva, 2002, 2003).”
3-D Depth Reconstruction from a Single Still Image (2007, background): “Torralba and Oliva [19] studied the Fourier spectrum of the images to compute the mean depth of a scene.”
BLINK: Multimodal Large Language Models Can See but Not Perceive (2024, background): “…optical properties like reflectance [12, 76], 3D primitives through multi-view reasoning [37,59], geometric reasoning through depth estimation [74], instance recognition through visual correspondence [55], affordance through keypoint grounding [36], and forensics through intrinsic images [9].”
Saliency Detection: A Spectral Residual Approach (2007, background): “This part, we believe is inherently responsible to the popping up of protoobjects in the pre-attentive stage.”
Depth Induced Feature Representation for 4D Human Activity Recognition (2014, methodology): “The depth constraints on the structure of the 3D Scenes and activities can be directly transposed into image/ video content [36].”
Contextual Priming for Object Detection (2003, background): “The organization is correlated with the size of the space that the scene subtends (Torralba and Oliva, 2002).”
Toward Holistic Scene Understanding: Feedback Enabled Cascaded Classification Models (2012, methodology): “To incorporate scene-level information, Torralba and Oliva [31], [32] used the statistics of low-level features across the entire scene to prime object detection or help depth estimation.”
Recovering Surface Layout from an Image (2007, methodology): “They further provide an algorithm for estimating these properties from the spectral signature of an image and, in [52], used similar algorithms to estimate mean depth of the scene.”
Towards Holistic Scene Understanding: Feedback Enabled Cascaded Classification Models (2010, methodology): “[31, 32] used the statistics of low-level features acros s the entire scene to prime object detection or help depth estimation.”
Modeling Search for People in 900 Scenes: A combined source model of eye guidance (2009, neutral): “Statistical regularities of natural scenes provide rich cues to target location and appearance (Eckstein, Drescher & Shimozaki, 2006; Hoiem, Efros, & Hebert, 2006; Oliva & Torralba, 2007; Torralba & Oliva, 2002, 2003).”