Feature detection (computer vision): Difference between revisions

Revision as of 16:16, 4 January 2016

Revision as of 21:11, 30 May 2016

Line 12:

There is no universal or exact definition of what constitutes a feature, and the exact definition often depends on the problem or the type of application. Given that, a feature is defined as an "interesting" part of an [[Digital image|image]], and features are used as a starting point for many computer vision algorithms. Since features are used as the starting point and main primitives for subsequent algorithms, the overall algorithm will often only be as good as its feature detector. Consequently, the desirable property for a feature detector is ''repeatability'': whether or not the same feature will be detected in two or more different images of the same scene.

Feature detection is a low-level [[image processing]] operation. That is, it is usually performed as the first operation on an image, and examines every [[pixel]] to see if there is a feature present at that pixel. If this is part of a larger algorithm, then the algorithm will typically only examine the image in the region of the features. As a built-in pre-requisite to feature detection, the input image is usually smoothed by a [[Gaussian blur|Gaussian]] kernel in a [[scale space|scale-space representation]] and one or several feature images are computed, often expressed in terms of local [[Image_DerivativesImage Derivatives|image derivatives]] operations.

Occasionally, when feature detection is [[computationally expensive]] and there are time constraints, a higher level algorithm may be used to guide the feature detection stage, so that only certain parts of the image are searched for features.

Line 142:

*{{Cite journal|last=Canny|first=J.|author-link=John Canny|title=A Computational Approach To Edge Detection|journal=IEEE Trans. Pattern Analysis and Machine Intelligence|volume=8|pages=679–714|year=1986|doi=10.1109/TPAMI.1986.4767851|issue=6}}. ([[Canny edge detector|Canny edge detection]])

* {{cite conference

| authorauthor1=C. Harris and |author2=M. Stephens | title=A combined corner and edge detector

| title=A combined corner and edge detector

| booktitle=Proceedings of the 4th Alvey Vision Conference

| pages=147–151

Line 150 ⟶ 149:

|format=PDF}}(Harris/Plessey corner detection)

*{{cite journal

| authorauthor1=S. M. Smith and |author2=J. M. Brady | title=SUSAN - a new approach to low level image processing

| title=SUSAN - a new approach to low level image processing

| url=http://citeseer.ist.psu.edu/smith95susan.html

| journal=International Journal of Computer Vision

Line 161 ⟶ 159:

}}(The SUSAN corner detector)

* {{cite conference

| authorauthor1=J. Shi and |author2=C. Tomasi | title=Good Features to Track,

| title=Good Features to Track,

| publisher=Springer

|date=June 1994

Line 169 ⟶ 166:

}}(The Shi and Tomasi corner detector)

*{{cite conference

| authorauthor1=E. Rosten and |author2=T. Drummond | title=Machine learning for high-speed corner detection

| title=Machine learning for high-speed corner detection

| booktitle=European Conference on Computer Vision

| publisher=Springer

Line 202 ⟶ 198:

}}(DOG blob detection with automatic scale selection)

*{{cite conference

| authorauthor1=J. Matas, |author2=O. Chum, |author3=M. Urban and |author4=T. Pajdla | title=Robust wide baseline stereo from maximally stable extremum regions

| title=Robust wide baseline stereo from maximally stable extremum regions

| booktitle=British Machine Vision Conference

| year=2002