1995
DOI: 10.1117/12.205303
|View full text |Cite
|
Sign up to set email alerts
|

Automatic and semiautomatic methods for image annotation and retrieval in query by image content (QBIC)

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
3
2

Citation Types

0
46
0

Year Published

1997
1997
2008
2008

Publication Types

Select...
5
2
1

Relationship

0
8

Authors

Journals

citations
Cited by 85 publications
(46 citation statements)
references
References 0 publications
0
46
0
Order By: Relevance
“…The polarity is a measure of the extent to which the gradient vectors in a certain neighborhood all point in the same direction. 2 (In the computation of second moments, this information is lost in the outer product operation; i.e., gradient vector directions differing by 180 are indistinguishable.) The polarity at a given pixel is computed with respect to the dominant orientation in the neighborhood of that pixel.…”
Section: Scale Selectionmentioning
confidence: 99%
See 2 more Smart Citations
“…The polarity is a measure of the extent to which the gradient vectors in a certain neighborhood all point in the same direction. 2 (In the computation of second moments, this information is lost in the outer product operation; i.e., gradient vector directions differing by 180 are indistinguishable.) The polarity at a given pixel is computed with respect to the dominant orientation in the neighborhood of that pixel.…”
Section: Scale Selectionmentioning
confidence: 99%
“…(1) is a sliding inner product, not a convolution, since x; y is spatially variant. 2 The polarity is related to the quadrature phase as discussed in [9,11]. where q + and q , are the rectified positive and negative parts of their argument,n is a unit vector perpendicular to , and Ω represents the neighborhood under consideration.…”
Section: Scale Selectionmentioning
confidence: 99%
See 1 more Smart Citation
“…Several other matching functions have been suggested in the literature [37], [42], [66]. Some of the most widespread techniques are: histogram matching [39]; texture matching [2]; intensity cross correlation [52]; optical flow matching [47]; kernel-based classification methods [17]; boosting classification methods [19], [44]; information divergence minimization [81], [77], [76], [29]; and mutual information (MI) maximization [84], [28], [53], [11]. The last two methods can be called "entropic methods" since both use a matching criterion defined as a relative entropy between the feature distributions.…”
Section: Introductionmentioning
confidence: 99%
“…Since the first appearance of multimedia databases, research on efficient means of feature extraction has become popular. To this end, some manual, semi-automatic and automatic tools have been developed [3,4,[6][7][8].…”
Section: Introductionmentioning
confidence: 99%