Contour-based learning for object detection

By Jamie Shotton, Andrew Blake, Roberto Cipolla. 2005

370 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition

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

Citations in the corpus

Pixelwise Instance Segmentation with a Dynamically Instantiated Network (2017, methodology): “Previous approaches to incorporating shape priors in segmentation [22, 7, 43] have involved generating “shape exemplars” from the training dataset and, at inference time, matching these exemplars to object proposals using the Chamfer distance [39, 31].”
DeepEdge: A multi-scale bifurcated deep network for top-down contour detection (2014, methodology): “Contour detection is typically considered a low-level problem, and used to aid higher-level tasks such as object detection [2, 3, 27, 21].”
Crisp Boundary Detection Using Pointwise Mutual Information (2014, methodology): “Object boundary cues have been used to aid in segmentation [1,2,3], object detection and recognition [4,5], and recovery of intrinsic scene properties such as shape, reflectance, and illumination [6].”
A Boundary-Fragment-Model for Object Detection (2006, methodology): “The principal differences are: the level of segmentation required in training ([24] requires more); the number of boundary fragments employed in each weak detector (a single fragment in [24], and a variable number here); and the method of localizing the detected centroid (grid in [24], mean shift he”
An Exemplar Model for Learning Object Classes (2007, methodology): “Though any previous model that requires manual bounding box annotation, such as [14, 18, 20], could now be trained automatically in this manner.”
Incremental learning of object detectors using a visual shape alphabet (2006, methodology): “0 [5],C [10],D [15],D [15],D [10],D I,T 7.”
ImageNet Auto-Annotation with Segmentation Propagation (2014, background): “Segmentation is useful in many higher-level applications such as object recognition, as it provides a spatial support for extracting texture and shape descriptors on objects [66,58].”
Cost-Sensitive Boosting (2011, background): “Both ideas have been explored in detail in the literature, with classifiers that explicitly segment the object to detect [51], [48], [55], [49], [56], learn configurations of its parts [53], [50] or both [48], [49].”
Sketch Tokens: A Learned Mid-level Representation for Contour and Object Detection (2013, background): “While numerous detection approaches [5, 11, 7] utilize such features directly, other papers learn edge-based features [27, 38, 21] or class-speciﬁc edges [23, 18] using object level supervision.”
Semantic part segmentation using compositional model combining shape and appearance (2014, background): “There are lots of works in the literature on modeling object shape such as [25, 15, 28, 20].”
Distance Transforms of Sampled Functions (2012, methodology): “For example, it was applied to study urban change in [15], motion planning in [25], object recognition in [29] and image analysis PDEs in [22].”
Discriminative Object Class Models of Appearance and Shape by Correlatons (2006, background): “Constellation models [20, 8], pictorial structures [12, 7] and fragment-based models [18, 16] have been proven powerful in that respect.”
Better Appearance Models for Pictorial Structures (2009, background): “human full bodies [8, 19], horses [25], sheep [9])”
A Survey of Recent Advances in Face Detection (2010, background): “[86] learned their object detectors with a boosting algorithm and their feature set consisted of a randomly chosen dictionary of contour fragments.”
Evolutionary Hough Games for coherent object detection (2012, neutral): “Since we are not obtaining bounding boxes but rather the sets of contributing voting elements for each person, we decided to evaluate the detection results with the strict criterion introduced in [30].”
Segmentation Propagation in ImageNet (2012, neutral): “Segmentation is useful in many higher-level applications such as object recognition, as it provides the spatial support for extracting texture and shape descriptors on objects [5, 6].”
Groups of Adjacent Contour Segments for Object Detection (2008, background): “In spite of their substantial scope, only comparably few works [1], [15], [24], [30] have tackled the class-level localization problem using contour features.”
Representing shape with a spatial pyramid kernel (2007, background): “Third, we show that shape and appearance kernels may be combined (again by learning parameters on a validation set).”
Human detection from images and videos: A survey (2016, background): “Studies in both human vision [5, 149, 130, 166] and computer vision [103, 142, 177] have indicated that edges of an object play an important role in object detection and recognition.”