Object Recognition in the Geometric Era: A Retrospective

By J. Mundy. 2006

136 citations recorded by the source corpus.

Subfields: Cognitive Neuroscience, Computer Vision and Pattern Recognition, Geometry and Topology

How researchers see this paper

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

In Cognitive Neuroscience

In Computer Vision and Pattern Recognition

In Geometry and Topology

Citations in the corpus

Lifting Object Detection Datasets into 3D (2015, background): “Modelbased recognition held important advantages, nevertheless [17]: modeling the 3D geometry of an object enabled arbi-”
Marr Revisited: 2D-3D Alignment via Surface Normal Prediction (2016, methodology): “Furthermore, there are stores of 3D models available for use in CAD repositories such as Trimble3D Warehouse and via capture from 3D sensor devices.”
Constrained parametric min-cuts for automatic object segmentation (2010, background): “As a result, in the nineties, part of the recognition community lost hope that a reliable segmentation procedure would be found and began investigating solutions that avoided botto mup segmentation altogether [20].”
Localizing 3D cuboids in single-view images (2012, background): “Traditional approaches have sought to recover 3D properties, such as creases, folds, and occlusions of surfaces, from a line representation extracted from the image [18].”
Semantic hierarchies for image annotation: A survey (2012, background): “Reviews of the state of the art in object recognition can be found in Mundy [68], for a historical perspective, Pinz [73] for a lengthy review of techniques used in object categorization, and Bosch et al.”
Seeing 3D Chairs: Exemplar Part-Based 2D-3D Alignment Using a Large Dataset of CAD Models (2014, background): “From its very beginnings [32] and up until the early nineties [29], object recognition research has been heavily geometry-centric.”
Reconstructing PASCAL VOC (2014, background): “Model-based recognition holds important advantages, nevertheless [20]: modeling the 3D geometry of an object enables arbitrary viewpoints and occlusion patterns to be rendered and recognized, and it also facilitates higher-level reasoning about interactions between objects and a scene.”
Boundary Cues for 3D Object Shape Recovery (2013, background): “As data-driven approaches to recognition became popular, researchers began to represent shape implicitly through weighted image gradient features, rather than explicitly through reconstruction [13].”
CPMC: Automatic Object Segmentation Using Constrained Parametric Min-Cuts (2012, background): “As a consequence, in 1990s, part of the recognition community lost confidence that a reliable segmentation procedure would be found and began investigating solutions that avoided bottom-up segmentation entirely [21].”
3D ShapeNets for 2.5D Object Recognition and Next-Best-View Prediction (2014, background): “Even though there are many theories about 3D representation [4, 18], the success of 3D-based methods is largely limited to instance recognition, using model-based keypoint matching [20, 24].”
Learning the semantics of object–action relations by observation (2011, methodology): “Geometry-based approaches use a geometric description of a 3D object and match its projected shape against the image of the object (Mundy and Zisserman 1992; Mundy 2006).”
Deep Exemplar 2D-3D Detection by Adapting from Real to Rendered Views (2015, background): “Due to the large appearance gap across the two domains, we learn how to adapt features of natural images to better match features for rendered views (Section 2).”
3 D ShapeNets : A Deep Representation for Volumetric Shape Modeling (2015, background): “Even though there are many theories about 3D representation [5, 21], the success of 3D-based methods has largely been limited to instance recognition, using model-based keypoint matching [24, 30].”