Scale &amp; Affine Invariant Interest Point Detectors

Mikolajczyk, Krystian; Schmid, Cordelia

doi:10.1023/b:visi.0000027790.02288.f2

Cited by 3,308 publications

(2,385 citation statements)

References 41 publications

Supporting

Mentioning

2,356

Contrasting

Unclassified

Order By: Relevance

“…The Pascal dataset also provides for each image a bounding box indicating the localization of the object. The local image descriptors were obtained by first using the Harris-Laplace detector [22] to extract interest points and by then using the SIFT descriptor [20] to represent the scale-invariant regions around these points. The dimension of the obtained SIFT features is 128.…”

Section: Bicyclementioning

confidence: 99%

Robust supervised classification with mixture models: Learning from data with uncertain labels

Bouveyron

Girard

2009

Pattern Recognition

119

View full text Add to dashboard Cite

. Robust supervised classification with mixture models: Learning from data with uncertain labels. Pattern Recognition, Elsevier, 2009, 42 (11), pp.2649-2658. <10.1016/j.patcog.2009 Robust supervised classification with mixture models:Learning from data with uncertain labels AbstractIn the supervised classification framework, human supervision is required for labeling a set of learning data which are then used for building the classifier. However, in many applications, human supervision is either imprecise, difficult or expensive. In this paper, the problem of learning a supervised multiclass classifier from data with uncertain labels is considered and a modelbased classification method is proposed to solve it. The idea of the proposed method is to confront an unsupervised modelling of the data with the supervised information carried by the labels of the learning data in order to detect inconsistencies. The method is able afterward to build a robust classifier taking into account the detected inconsistencies into the labels. Experiments on artificial and real data are provided to highlight the main features of the proposed method as well as an application to object recognition under weak supervision.

show abstract

Section: Bicyclementioning

confidence: 99%

Robust supervised classification with mixture models: Learning from data with uncertain labels

Bouveyron

Girard

2009

Pattern Recognition

119

View full text Add to dashboard Cite

show abstract

“…The test image set consists of real structured and textured images of various scenes, with different geometric and photometric transformations such as viewpoint change, image blur, illumination change, scale and rotation and image compression. For the detectors presented here we describe a circular region with a diameter that is 3× the detected scale of the point of interest, similar to the approach in [10,11]. The overlap of the circular regions corresponding to an interest point pair in a set of images is measured based on the ratio of intersection and union of the circular regions.…”

Section: Resultsmentioning

confidence: 99%

“…Thus, a set of features may be detected at multiple scales. However, applying a detector at multiple scales may introduce other issues, as the same feature may be present over a range of scales within the detector's range [10], and by representing the same feature at many scales we increase the difficulty of matching the detected features. Hence, a scale invariant approach is more appropriate, where the characteristic scale of the feature is identified.…”

Section: Introductionmentioning

confidence: 99%

“…Hence, a scale invariant approach is more appropriate, where the characteristic scale of the feature is identified. This characteristic scale is the scale at which the feature is most strongly represented, and it is not related to the resolution of the image, but rather the underlying structure of the detected feature [10]. By using an operator to measure the response of the same interest point at different scales, the scale at which the peak response is obtained can be identified.…”

Section: Introductionmentioning

confidence: 99%

“…Performance is evaluated with respect to repeatability and feature matching using the evaluation techniques presented in [10], highlighting improvements when compared with other well known interest point detectors and descriptors.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

A Finite Element Blob Detector for Robust Features

Kerr

Coleman

Scotney

2011

Image Analysis and Processing – ICIAP 2011

View full text Add to dashboard Cite

Abstract. Traditionally feature extraction is focussed on edge and corner detection, however, more recently points of interest and blob like features have also become prominent in the field of computer vision and are typically used to determine correspondences between two images of the same scene. We present a new approach to a Hessian blob detector, designed within the finite element framework, which is similar to the multi-scale approach applied in the SURF detector. We present performance evaluation that demonstrates the accuracy of our approach in comparison to well known existing algorithms.

show abstract

Near‐Duplicate Image and Video Detection

Zhao

Ngo

2015

Wiley Encyclopedia of Electrical and Electronics Engineering

View full text Add to dashboard Cite

As the bandwidth accessible to average users is increasing, multimedia data, in particular images and videos, become the fastest growing data type on Internet. Especially with the popularity of social media, there has been exponential growth in images and videos available on the Web. On one hand, among these huge volumes of images and videos, there exist large numbers of near‐duplicates and copies. Near‐duplicates carry both blessing and redundant signals. For example, it provides rich visual clue for indexing and summarizing broadcast videos from different channels. On the other hand, the excessive amount of near‐duplicates makes browsing web videos streamed over Internet an extremely time‐consuming task. In this paper, we provide an overview of the state‐of‐the‐art researches on the algorithms and techniques for large‐scale mining of visual near‐duplicates. It basically covers issues about image representation, image feature indexing and matching, pattern modeling, scalable detection, and emerging multimedia applications that are built upon near‐duplicate detection.

show abstract

Scale & Affine Invariant Interest Point Detectors

Cited by 3,308 publications

References 41 publications

Robust supervised classification with mixture models: Learning from data with uncertain labels

Robust supervised classification with mixture models: Learning from data with uncertain labels

A Finite Element Blob Detector for Robust Features

Near‐Duplicate Image and Video Detection

Contact Info

Product

Resources

About