Top Banner

Click here to load reader

Specific Object Recognition using SIFT

Feb 22, 2016

ReportDownload

Documents

bryson

Machine Vision and Image Processing Group (Student Group) Electronic Research Center of Iran University of Science and Technology http://mvip.iust.ac.ir. Specific Object Recognition using SIFT. Presentation by: Amir Azizi. گروه بینایی ماشین و پردازش تصویر. آبان 1389 November 2010. - PowerPoint PPT Presentation

Specific Object Recognition using SIFT

Specific Object Recognition using SIFT Machine Vision and Image Processing Group (Student Group)Electronic Research Center of Iran University of Science and Technologyhttp://mvip.iust.ac.irPresentation by:

Amir Azizi 1389November [email protected]

Introduction

Example for specific object recognition:Search photos on web for the particular placesj. [email protected]

IntroductionWhy is it difficult?

j. [email protected]

Challenges

1- [email protected]

Challenges2- Illumination

[email protected]

Challenges3- Occlusion

[email protected]

Challenges4- Scale

[email protected]

Challenges5- Deformation

[email protected]

Challenges6- Background Clutter

[email protected]

Local Features

NewDatasetCornersBlobsLocal Features (Interest points or key points):Some of applications:Specific object recognitionTrackingImage registrationCamera [email protected]

Local FeaturesDesired Properties of local features:Repeatability- The same feature can be found in several images despite geometric and photometric transformationDistinctiveness- Each feature has a distinctive descriptionLocality- A feature occupies a relatively small area of the image; robust to clutter and occlusionQuantity- Number of featuresEfficiency- Applications that need to [email protected]

Local feature-based object recognition :

:[email protected]

HarrisHarris-LaplaceHarris-AffineMSERSalient RegionsSIFT Detector (DoG)SURF DetectorHessian-LaplaceHessian-AffineShape ContextGeometric BlurSIFT DescriptorSURF DescriptorLocal feature-based object [email protected]

SIFT

SIFT: Scale Invariant Feature Transform1999 and [email protected]

Hessian MatrixIn mathematics, the Hessian matrix (or simply the Hessian) is the square matrix of second-order partial derivatives of a function; that is, it describes the local curvature of a function of many variables.We want to find Blobs, soSIFT uses extrema of Hessian matrix trace:[email protected]

1- SIFT Detector : .1- Lindeberg 1994,19982- Koendernik 1984SCALE-SPACEScale = [email protected]

1- SIFT Detector SCALE-SPACEDoGMikolajczyk 2002: normalized Laplacian gives more robust features [email protected]

1- SIFT Detector

Down sampling :[email protected]

1- SIFT Detector

:[email protected]

1- SIFT Detector :

[email protected]

1- SIFT Detector :

[email protected]

2- SIFT Descriptor

[email protected]

2- SIFT Descriptor

Rotation Invariance:[email protected]

2- SIFT Descriptor

So we have a feature vector with 128 [email protected]

3- Matching* k-d k * BBF *

David Lowe * .25

[email protected]