ECE 598: An Invitation to 3-D Vision: From Images to Geometric Models

By Y. Ma, Stefano Soatto, Jana Koseck, S. Sastry. 2004

1,185 citations recorded by the source corpus.

View Paper: https://www.semanticscholar.org/paper/dae123620f0949634981884a33a93d26bedef077

How researchers see this paper

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

Citations in the corpus

Energy-Based Geometric Multi-model Fitting (2012, background): “This fundamental matrix satisfies the epipolar constraint p T r Fp l = 0 where p r and p l are two matching features corresponding to a 3D point on some rigid body (Ma et al. 2003; Li 2007).”
Vision based navigation for micro helicopters (2012, methodology): “Geometrically, the intersection of two such planes unambiguously defines the direction of ~v (Ma et al. (2000)).”
Looking at Vehicles on the Road: A Survey of Vision-Based Vehicle Detection, Tracking, and Behavior Analysis (2013, methodology): “A visual object detection system requires camera-based sensing to measure the scene’s light, as well as computational machinery to extract information from raw image data [12].”
A Unifying Contrast Maximization Framework for Event Cameras, with Applications to Motion, Depth, and Optical Flow Estimation (2018, background): “The parameters v / d account for the well-known fact that, without additional information, there is a scale ambiguity: the decomposition of a planar homography [35, 39] only provides the direction of the translation, but not its magnitude.”
Vehicle Detection Techniques for Collision Avoidance Systems: A Review (2015, background): “The variation in left and right images between corresponding pixels is called disparity and stereo matching of disparities of all image points calculates the disparity map [113].”
Visual-inertial navigation, mapping and localization: A scalable real-time causal approach (2011, methodology): “Most work that relies on the traditional epipolar geometry pipeline (Ma et al. 2003), including various sorts of bundle adjustment, introduces delays in visual processing, because one has to wait for a small batch of images to be collected that have ‘sufficient parallax’.”
Particle Filter SLAM with High Dimensional Vehicle Model (2009, methodology): “These feature points are then searched for in the subsequent images using cross-correlations [25] according to Algorithm 2.”
Research Advances and Challenges of Autonomous and Connected Ground Vehicles (2019, background): “Stereo matching is a transformation, which is applied to build the disparity map for the stereo camera [151].”
Visual servo control. I. Basic approaches (2006, background): “The details of imaging geometry and perspective projection can be found in many computer vision texts, including [3], [4].”
Dark, Beyond Deep: A Paradigm Shift to Cognitive AI with Humanlike Common Sense (2020, background): “construction from a single 2D image is a well-known ill-posed problem; there may exist an infinite number of possible 3D configurations that match the projected 2D observed images [30].”
Real-time onboard visual-inertial state estimation and self-calibration of MAVs in unknown environments (2012, methodology): “In [14] the authors use the solution of (6) to find a globally and temporally consistent scale factor for subsequent camera frames (i.”
SVO: Semidirect Visual Odometry for Monocular and Multicamera Systems (2017, background): “The position and orientation of the world frame W with respect to the kth camera frame is described by the rigid body transformation TkW ∈ SE(3) [47].”
SVO: Fast semi-direct monocular visual odometry (2014, background): “The twist coordinates ξ are mapped to SE(3) by the exponential map [26]:”
Visual-Inertial Sensor Fusion: Localization, Mapping and Sensor-to-Sensor Self-calibration (2011, background): “…1 ] = K [ xi/zi yi/zi 1 ] , (23) where xi/zi and yi/zi are the normalized image plane coordinates, ui and vi are the observed pixel coordinates, K is the 3 × 3 camera intrinsic calibration matrix (Ma et al. 2004), and ηi is a Gaussian measurement noise vector with covariance matrix Ri = σ 2i I2.”