Evaluation of Interest Point Detectors

By Cordelia Schmid, Roger Mohr, Christian Bauckhage. International Journal of Computer Vision, 2000

1,787 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/2713e7a59105a832e20c01c3c202b9dcd2b5f889

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

An automatic image registration for applications in remote sensing (2005, methodology): “In this implementation, the CPs are detected in the reference image by applying the improved version of the Harris corner detector at regions where the gradient magnitude is high.”
Machine Learning for High-Speed Corner Detection (2006, methodology): “When measuring reliability, what is important is if the same real-world features are detected from multiple views [1] This is the definition which will be used here.”
Simultaneous Localization and Mapping: A Survey of Current Trends in Autonomous Driving (2017, neutral): “As such, sensor data extraction, primitive search [9], data association [10] or map storage and updates [11] are part of the topics which concern SLAM as well.”
Closed-Loop Learning of Visual Control Policies (2011, neutral): “Practical experiments on simulated visual navigation tasks exhibit this deficiency, as soon as the number of detected visual features is reduced or as features are made more similar by using a less sensitive metric.”
Scalable Recognition with a Vocabulary Tree (2006, background): “represent an object with descriptors extracted from these local regions [1, 2, 8, 11, 15, 16, 18].”
Invariant Features from Interest Point Groups (2002, background): “Repeatibility is defined for two images related by a homography as the fraction of interest points which are consistent with that homography, up to some tolerance (see [11]).”
SAR-SIFT: A SIFT-Like Algorithm for SAR Images (2015, background): “The repeatability criterion [41] gives a measure of the stability of keypoint detection, with regard to the image changes.”
Diagnosing Error in Object Detectors (2012, methodology): “, [13, 14]), contextual methods [15, 16], dataset design [17], and cross-dataset generalization [18].”
Linear Time Maximally Stable Extremal Regions (2008, background): “Extraction of invariant regions has recently been the focus of intense study [1,2,3,4,5,6,7,8,9,10,11,12], supporting a wide variety of applications such as recognition, image retrieval, mosaicing, 3D reconstruction, tracking, robot navigation and more.”
Multi-image matching using multi-scale oriented patches (2005, methodology): “There has also been good progress in evaluating these techniques with respect to interest point repeatability [12] and descriptor performance [13].”
Generic object recognition with boosting (2006, methodology): “Based on the evaluation of interest point detectors by Schmid et al. [31], we decided to use the scale invariant Harris-Laplace detector [26] and the affine invariant interest point detector [27], both by Mikolajczyk and Schmid.”
Unsupervised Learning of Human Action Categories Using Spatial-Temporal Words (2008, methodology): “In spite of the existence of a fairly large variety of methods to extract interest points from static images (Schmid et al. 2000), less work has been done on space-time interest point detection in videos.”
Learning Sparse Representations for Human Action Recognition (2012, methodology): “This 2D keypoint detector has been shown to outperform other keypoint detectors in terms of distinctiveness and stability of the detected points under various image transformations like rotation, illumination variation, scale and viewpoint changes [28].”
Recognition with local features: the kernel recipe (2003, methodology): “Detection of interest points was done using a standard Harris-type corner detector, which was shown to have high repeatability and robust performance [23].”
Scale & Affine Invariant Interest Point Detectors (2004, neutral): “However , the Harris interest point detector is not invariant to scale and affine transformations (Schmid et al., 2000).”