DETECTING LEVELLING RODS
USING SIFT FEATURE MATCHING
GROUP 1
MSc Course 2006-08
25TH
June 2007
Sajid Pareeth
Sonam Tashi
Gabriel Vincent Sanya
Michael Mutale
PHOTOGRAMMETRY STUDIO
Objective
Introduction
SIFT Algorithm
SIFT-Keypoints
Extraction
Keypoints
Matching
Work Flow
RANSAC
Advantages/
Limitations
Demo
 To develop a MATLAB procedure for the detection of
levelling rods.
 To carry out matching based on Algorithms proposed by
David Lowe
 Object recognition using invariant features
Transformation
Examples
Summary of Steps
Images
SIFT Key Feature Extraction
Keypoint Matching
Removing Outliers - RANSAC
Transformation
Introduction
SIFT Algorithm
SIFT-Keypoints
Extraction
Keypoints
Matching
Work Flow
RANSAC
Advantages/
Limitations
Demo
Transformation
Examples
Background - SIFT
 Published by David Lowe et al. 1999-2004
 Algorithm to extract features that are invariant to
rotation,
scaling and
 partially invariant to
changes in illumination and camera viewpoint
 Resulting features are highly distinctive
Consists of (Major stages of computation):
 Scale-space extrema detection-the Gaussian Widow
 Keypoint localization- 4x4 samples per window in 8
directions
 Orientation assignment
 Keypoint descriptor-feature vector is modified to reduce the effects of illumination
change. The vector is normalized to unit length. large gradient magnitudes by thresholding the
values in the unit feature vector to each be no larger than 0.2, and then renormalizing to unit
length.
SIFT Algorithm
SIFT-Keypoints
Extraction
Keypoints
Matching
Work Flow
RANSAC
Advantages/
Limitations
Demo
Transformation
Examples
Introduction
A keypoint descriptor is created by first computing the
gradient magnitude and orientation at each image sample
point in a region around the keypoint location,
These are weighted by a Gaussian window, indicated by
the overlaid circle. These samples are then accumulated
into orientation histograms summarizing the contents over
4x4 subregions,
The length of each arrow corresponding to the sum of the
gradient magnitudes near that direction within
the region.
Scale Invariant Detection
 Consider regions (e.g. circles) of different sizes around
a point
 Regions of corresponding sizes will look the same in
both images SIFT-Keypoints
Extraction
Keypoints
Matching
Work Flow
RANSAC
Advantages/
Limitations
Demo
Transformation
Examples
SIFT Algorithm
Introduction
 The problem: how do we choose
corresponding circles independently in each
image?
Scale Invariant Detection
SIFT Algorithm
SIFT-Keypoints
Extraction
Keypoints
Matching
Work Flow
RANSAC
Advantages/
Limitations
Demo
Transformation
Examples
SIFT Algorithm
Introduction
 Solution:
 Design a function on the region (circle), which is
“scale invariant” (the same for corresponding
regions, even if they are at different scales)
Example: average intensity. For corresponding regions
(even of different sizes) it will be the same.
 For a point in one image, we can consider it as a
function of region size (circle radius)
scale = 1/2
f
region size
Image 1 f
region size
Image 2
Scale Invariant Detection
SIFT-Keypoints
Extraction
Keypoints
Matching
Work Flow
RANSAC
Advantages/
Limitations
Demo
Transformation
Examples
SIFT Algorithm
Introduction
 Common approach
scale = 1/2
f
region size
Image 1 f
region size
Image 2
Take a local maximum of this function
Observation: region size, for which the maximum is achieved,
should be invariant to image scale.
s1 s2
Important: this scale invariant region size is found
in each image independently!
Scale Invariant Detection
SIFT-Keypoints
Extraction
Keypoints
Matching
Work Flow
RANSAC
Advantages/
Limitations
Demo
Transformation
Examples
SIFT Algorithm
Introduction
scale
x
y
← DoG →
←DoG→
SIFT(Lowe)
Maxima and minima of the difference-of-Gaussian
images detected by comparing a pixel (marked with X)
to its 26 neighbors in 3x3 regions at the current and
adjacent scales (marked with circles).
SIFT-Keypoints
Extraction
Keypoints
Matching
Work Flow
RANSAC
Advantages/
Limitations
Demo
Transformation
Examples
SIFT Algorithm
Introduction
Gaussian pyramids
Scale
First Octave
Scale
Next
Octave
Difference of Gaussians
Difference of Gaussian(DoG)
SIFT-Keypoints
Extraction
Keypoints
Matching
Work Flow
RANSAC
Advantages/
Limitations
Demo
Transformation
Examples
SIFT Algorithm
Introduction
Keypoint Descriptors
Feature Vectors
 Thresholded image gradients sampled over:
 16x16 array of locations in scale space
 Histogram of 4x4 samples per window in 8 directions
 Gaussian weighting around center
 8 orientations x 4 x 4 histogram array = 128 dimensional
feature vector
 4x4 Gradient window
SIFT Algorithm
Keypoints
Matching
Work Flow
RANSAC
Advantages/
Limitations
Demo
Transformation
Examples
SIFT-Keypoints
Extraction
Introduction
Gaussian Gradient Window
Criteria for Selection
 Prominent
 Distinguishable-from neighborhood
 Invariant
 Stable to disturbances
 Rare (exceptional) - from other selected points
 Meaningful - with respect to image interpretation
SIFT Algorithm
Keypoints
Matching
Work Flow
RANSAC
Advantages/
Limitations
Demo
Transformation
Examples
SIFT-Keypoints
Extraction
Introduction
SIFT keypoints
 Detected Keypoints in reference and candidate image
 Keypoints in both images will be matched
Candidate image with keypointsReference image
with keypoints
SIFT Algorithm
Keypoints
Matching
Work Flow
RANSAC
Advantages/
Limitations
Demo
Transformation
Examples
SIFT-Keypoints
Extraction
Introduction
Keypoint Matching
 Fundamental aspect in computer vision.
 Based on Euclidean distance of the feature vectors.
 Nearest neighbor algorithms.
 Product between descriptors is calculated.
 Inverse cosine of the products gives the Euclidean distance
 Matches with Ratio of vector angles from the nearest to second
nearest neighbor less than distRatio value are selected.
SIFT Algorithm
Keypoints
Matching
Work Flow
RANSAC
Advantages/
Limitations
Demo
Transformation
Examples
SIFT-Keypoints
Extraction
Introduction
90% of the false matches are
removed.
False matches due to ambiguos
features or features arise from
background clutter.
Reliable object recognition with
few best Matches
Remove outliers
Matched points
SIFT Algorithm
SIFT-Keypoints
Extraction
Work Flow
RANSAC
Advantages/
Limitations
Demo
Transformation
Examples
Keypoints
Matching
Introduction
RANSAC: Algorithm
RANdom SAmple Consensus
 Estimate parameters of a mathematical model from a set of
observed data which contains outliers.
 A model is fitted using hypothetical inliers.
 If other data fits to the model, added to the inliers.
 Reestimated the model with the new set of inliers.
 We used ransac-fit-homography which Robustly fits a
homography to a set of matched points.
SIFT Algorithm
SIFT-Keypoints
Extraction
Work Flow
Advantages/
Limitations
Demo
Transformation
Examples
Keypoints
Matching
RANSAC
Introduction
Ransac
RANSAC: Result
SIFT Algorithm
SIFT-Keypoints
Extraction
Work Flow
RANSAC
Advantages/
Limitations
Demo
Transformation
Examples
Keypoints
Matching
Introduction
Make the shortest image the
same height as the other image.
Append Images
Transformation
tform = cp2tform(M1,M2,’affine')
SIFT Algorithm
SIFT-Keypoints
Extraction
Keypoints
Matching
Work Flow
Advantages/
Limitations
Demo
Transformation
Examples
RANSAC
Introduction
TStructure
2
3
Rod vs. OpenShrubs
1
Matches: 9
2
Rod: 626 keypoints found
Image: 28958 keypoints found
Matches: 16
…Good Matches
SIFT Algorithm
SIFT-Keypoints
Extraction
Keypoints
Matching
Work Flow
RANSAC
Advantages/
Limitations
Demo
Examples
Transformation
Introduction
2
Rod Vs Human Occlusion??
1
Rod: 626 keypoints found
Image: 2347 keypoints found
Matches: 155
Rod: 626 keypoints found
Image: 2347 keypoints found
Matches: 144
2
3
…Good Matches
SIFT Algorithm
SIFT-Keypoints
Extraction
Keypoints
Matching
Work Flow
RANSAC
Advantages/
Limitations
Demo
Transformation
Examples
Introduction
Limitations
1
Rod: 626 keypoints found
Image: 17927 keypoints found
Matches: 16
Matches: 4
2
Rod: 626 keypoints found
Image: 30505 keypoints found
3
SIFT Algorithm
SIFT-Keypoints
Extraction
Keypoints
Matching
Work Flow
RANSAC
Advantages/
Limitations
Demo
Transformation
Examples
Introduction
Advantages
 Locality: features are local, so robust to occlusion and clutter
(no prior segmentation)
 Distinctiveness: individual features can be matched to a large
database of objects
 Quantity: many features can be generated for even small
objects
 Efficiency: close to real-time performance
 Extensibility: can easily be extended to wide range of differing
feature types, with each adding robustness
SIFT Algorithm
SIFT-Keypoints
Extraction
Keypoints
Matching
Work Flow
RANSAC
Advantages/
Limitations
Demo
Transformation
Examples
Introduction
Problems/Enhancement
 Only invariant to affine transformations to a certain degree
 Best performance on highly textured images
 Use Principal components analysis PCA instead of Gaussian
weighting for gradients
Demo
SIFT Algorithm
SIFT-Keypoints
Extraction
Keypoints
Matching
Work Flow
RANSAC
Advantages/
Limitations
Transformation
Examples
Introduction
Demo
Demo
SIFT Algorithm
SIFT-Keypoints
Extraction
Keypoints
Matching
Work Flow
RANSAC
Advantages/
Limitations
Transformation
Examples
Introduction
THANK YOU!!

More Related Content

PPTX
Computer vision - edge detection
PPTX
Scale Invariant feature transform
PPTX
Spatial Filters (Digital Image Processing)
PPTX
Deep Learning in Bio-Medical Imaging
PPTX
Moving object detection
PPTX
Deep learning for image super resolution
PDF
Lecture 11
Computer vision - edge detection
Scale Invariant feature transform
Spatial Filters (Digital Image Processing)
Deep Learning in Bio-Medical Imaging
Moving object detection
Deep learning for image super resolution
Lecture 11

What's hot (20)

PDF
Convolutional Neural Networks (CNN)
PPTX
Computer Vision Introduction
PDF
ViT (Vision Transformer) Review [CDM]
PPTX
AlexNet
PPTX
Transformer in Vision
PDF
Depth estimation do we need to throw old things away
PPTX
Tutorial on Object Detection (Faster R-CNN)
PDF
Image Recontruction
PPT
Basics of edge detection and forier transform
PPTX
Edge detection
PPTX
Snakes in Images (Active contour tutorial)
PDF
PR-302: NeRF: Representing Scenes as Neural Radiance Fields for View Synthesis
PPTX
Transformers in Vision: From Zero to Hero
PPTX
SPATIAL FILTERING IN IMAGE PROCESSING
PPTX
Canny Edge Detection
PPTX
Image Segmentation Using Deep Learning : A survey
PDF
Photo-realistic Single Image Super-resolution using a Generative Adversarial ...
PDF
Slides for Ph.D. Thesis Defense of Dheryta Jaisinghani at IIIT-Delhi, INDIA
PDF
Recurrent Neural Networks (RNN) | RNN LSTM | Deep Learning Tutorial | Tensorf...
PPTX
Kalman Filter | Statistics
Convolutional Neural Networks (CNN)
Computer Vision Introduction
ViT (Vision Transformer) Review [CDM]
AlexNet
Transformer in Vision
Depth estimation do we need to throw old things away
Tutorial on Object Detection (Faster R-CNN)
Image Recontruction
Basics of edge detection and forier transform
Edge detection
Snakes in Images (Active contour tutorial)
PR-302: NeRF: Representing Scenes as Neural Radiance Fields for View Synthesis
Transformers in Vision: From Zero to Hero
SPATIAL FILTERING IN IMAGE PROCESSING
Canny Edge Detection
Image Segmentation Using Deep Learning : A survey
Photo-realistic Single Image Super-resolution using a Generative Adversarial ...
Slides for Ph.D. Thesis Defense of Dheryta Jaisinghani at IIIT-Delhi, INDIA
Recurrent Neural Networks (RNN) | RNN LSTM | Deep Learning Tutorial | Tensorf...
Kalman Filter | Statistics
Ad

Similar to Feature Matching using SIFT algorithm (20)

PDF
Video Stitching using Improved RANSAC and SIFT
PDF
SHORT LISTING LIKELY IMAGES USING PROPOSED MODIFIED-SIFT TOGETHER WITH CONVEN...
PDF
SHORT LISTING LIKELY IMAGES USING PROPOSED MODIFIED-SIFT TOGETHER WITH CONVEN...
PPT
PCA-SIFT: A More Distinctive Representation for Local Image Descriptors
PDF
Svd filtered temporal usage clustering
PPTX
Realtime pothole detection system using improved CNN Models
PPTX
All projects
PPTX
Graphics
PDF
CNN vs SIFT-based Visual Localization - Laura Leal-Taixé - UPC Barcelona 2018
DOCX
Sift detector boosted by adaptive contrast threshold to improve matching robu...
DOCX
Sift detector boosted by adaptive contrast threshold to improve matching robu...
PDF
An automatic algorithm for object recognition and detection based on asift ke...
PDF
Lecture 01 frank dellaert - 3 d reconstruction and mapping: a factor graph ...
PPTX
casestudy_important.pptx
PDF
Intelligent Auto Horn System Using Artificial Intelligence
PPT
Using FME for Topographical Data Generalization at Natural Resources Canada
PPTX
EGUE Technikrom Final_8_12_13
PDF
Panoramic Imaging using SIFT and SURF
PPTX
Project Based on MATLAB For Final Year Research Guidance
PDF
Improved Characters Feature Extraction and Matching Algorithm Based on SIFT
Video Stitching using Improved RANSAC and SIFT
SHORT LISTING LIKELY IMAGES USING PROPOSED MODIFIED-SIFT TOGETHER WITH CONVEN...
SHORT LISTING LIKELY IMAGES USING PROPOSED MODIFIED-SIFT TOGETHER WITH CONVEN...
PCA-SIFT: A More Distinctive Representation for Local Image Descriptors
Svd filtered temporal usage clustering
Realtime pothole detection system using improved CNN Models
All projects
Graphics
CNN vs SIFT-based Visual Localization - Laura Leal-Taixé - UPC Barcelona 2018
Sift detector boosted by adaptive contrast threshold to improve matching robu...
Sift detector boosted by adaptive contrast threshold to improve matching robu...
An automatic algorithm for object recognition and detection based on asift ke...
Lecture 01 frank dellaert - 3 d reconstruction and mapping: a factor graph ...
casestudy_important.pptx
Intelligent Auto Horn System Using Artificial Intelligence
Using FME for Topographical Data Generalization at Natural Resources Canada
EGUE Technikrom Final_8_12_13
Panoramic Imaging using SIFT and SURF
Project Based on MATLAB For Final Year Research Guidance
Improved Characters Feature Extraction and Matching Algorithm Based on SIFT
Ad

More from Sajid Pareeth (6)

PDF
PhD defence presentation, 12 July 2016 @ FU-Berlin
PDF
Monitoring and retrieving historical daily surface temperature of sub-alpine ...
PDF
Inter-sensor comparison of lake surface temperatures derived from MODIS, AVHR...
PDF
An open source framework for processing daily satellite images (AVHRR) over l...
PPT
Spatial Data Infrastructure in Europe
PPTX
Irrigated Area Mapping, South Asia
PhD defence presentation, 12 July 2016 @ FU-Berlin
Monitoring and retrieving historical daily surface temperature of sub-alpine ...
Inter-sensor comparison of lake surface temperatures derived from MODIS, AVHR...
An open source framework for processing daily satellite images (AVHRR) over l...
Spatial Data Infrastructure in Europe
Irrigated Area Mapping, South Asia

Recently uploaded (20)

PPTX
MicrosoftCybserSecurityReferenceArchitecture-April-2025.pptx
PPT
Module 1.ppt Iot fundamentals and Architecture
PDF
sbt 2.0: go big (Scala Days 2025 edition)
PDF
Five Habits of High-Impact Board Members
PDF
Consumable AI The What, Why & How for Small Teams.pdf
PPTX
2018-HIPAA-Renewal-Training for executives
PPT
Galois Field Theory of Risk: A Perspective, Protocol, and Mathematical Backgr...
PDF
Developing a website for English-speaking practice to English as a foreign la...
PPTX
Benefits of Physical activity for teenagers.pptx
PDF
Taming the Chaos: How to Turn Unstructured Data into Decisions
PDF
Convolutional neural network based encoder-decoder for efficient real-time ob...
PDF
STKI Israel Market Study 2025 version august
PDF
A review of recent deep learning applications in wood surface defect identifi...
PDF
A Late Bloomer's Guide to GenAI: Ethics, Bias, and Effective Prompting - Boha...
PDF
Hindi spoken digit analysis for native and non-native speakers
PDF
Enhancing emotion recognition model for a student engagement use case through...
PDF
Abstractive summarization using multilingual text-to-text transfer transforme...
PDF
sustainability-14-14877-v2.pddhzftheheeeee
PDF
Architecture types and enterprise applications.pdf
PDF
Two-dimensional Klein-Gordon and Sine-Gordon numerical solutions based on dee...
MicrosoftCybserSecurityReferenceArchitecture-April-2025.pptx
Module 1.ppt Iot fundamentals and Architecture
sbt 2.0: go big (Scala Days 2025 edition)
Five Habits of High-Impact Board Members
Consumable AI The What, Why & How for Small Teams.pdf
2018-HIPAA-Renewal-Training for executives
Galois Field Theory of Risk: A Perspective, Protocol, and Mathematical Backgr...
Developing a website for English-speaking practice to English as a foreign la...
Benefits of Physical activity for teenagers.pptx
Taming the Chaos: How to Turn Unstructured Data into Decisions
Convolutional neural network based encoder-decoder for efficient real-time ob...
STKI Israel Market Study 2025 version august
A review of recent deep learning applications in wood surface defect identifi...
A Late Bloomer's Guide to GenAI: Ethics, Bias, and Effective Prompting - Boha...
Hindi spoken digit analysis for native and non-native speakers
Enhancing emotion recognition model for a student engagement use case through...
Abstractive summarization using multilingual text-to-text transfer transforme...
sustainability-14-14877-v2.pddhzftheheeeee
Architecture types and enterprise applications.pdf
Two-dimensional Klein-Gordon and Sine-Gordon numerical solutions based on dee...

Feature Matching using SIFT algorithm

  • 1. DETECTING LEVELLING RODS USING SIFT FEATURE MATCHING GROUP 1 MSc Course 2006-08 25TH June 2007 Sajid Pareeth Sonam Tashi Gabriel Vincent Sanya Michael Mutale PHOTOGRAMMETRY STUDIO
  • 2. Objective Introduction SIFT Algorithm SIFT-Keypoints Extraction Keypoints Matching Work Flow RANSAC Advantages/ Limitations Demo  To develop a MATLAB procedure for the detection of levelling rods.  To carry out matching based on Algorithms proposed by David Lowe  Object recognition using invariant features Transformation Examples
  • 3. Summary of Steps Images SIFT Key Feature Extraction Keypoint Matching Removing Outliers - RANSAC Transformation Introduction SIFT Algorithm SIFT-Keypoints Extraction Keypoints Matching Work Flow RANSAC Advantages/ Limitations Demo Transformation Examples
  • 4. Background - SIFT  Published by David Lowe et al. 1999-2004  Algorithm to extract features that are invariant to rotation, scaling and  partially invariant to changes in illumination and camera viewpoint  Resulting features are highly distinctive Consists of (Major stages of computation):  Scale-space extrema detection-the Gaussian Widow  Keypoint localization- 4x4 samples per window in 8 directions  Orientation assignment  Keypoint descriptor-feature vector is modified to reduce the effects of illumination change. The vector is normalized to unit length. large gradient magnitudes by thresholding the values in the unit feature vector to each be no larger than 0.2, and then renormalizing to unit length. SIFT Algorithm SIFT-Keypoints Extraction Keypoints Matching Work Flow RANSAC Advantages/ Limitations Demo Transformation Examples Introduction
  • 5. A keypoint descriptor is created by first computing the gradient magnitude and orientation at each image sample point in a region around the keypoint location, These are weighted by a Gaussian window, indicated by the overlaid circle. These samples are then accumulated into orientation histograms summarizing the contents over 4x4 subregions, The length of each arrow corresponding to the sum of the gradient magnitudes near that direction within the region.
  • 6. Scale Invariant Detection  Consider regions (e.g. circles) of different sizes around a point  Regions of corresponding sizes will look the same in both images SIFT-Keypoints Extraction Keypoints Matching Work Flow RANSAC Advantages/ Limitations Demo Transformation Examples SIFT Algorithm Introduction
  • 7.  The problem: how do we choose corresponding circles independently in each image? Scale Invariant Detection SIFT Algorithm SIFT-Keypoints Extraction Keypoints Matching Work Flow RANSAC Advantages/ Limitations Demo Transformation Examples SIFT Algorithm Introduction
  • 8.  Solution:  Design a function on the region (circle), which is “scale invariant” (the same for corresponding regions, even if they are at different scales) Example: average intensity. For corresponding regions (even of different sizes) it will be the same.  For a point in one image, we can consider it as a function of region size (circle radius) scale = 1/2 f region size Image 1 f region size Image 2 Scale Invariant Detection SIFT-Keypoints Extraction Keypoints Matching Work Flow RANSAC Advantages/ Limitations Demo Transformation Examples SIFT Algorithm Introduction
  • 9.  Common approach scale = 1/2 f region size Image 1 f region size Image 2 Take a local maximum of this function Observation: region size, for which the maximum is achieved, should be invariant to image scale. s1 s2 Important: this scale invariant region size is found in each image independently! Scale Invariant Detection SIFT-Keypoints Extraction Keypoints Matching Work Flow RANSAC Advantages/ Limitations Demo Transformation Examples SIFT Algorithm Introduction
  • 10. scale x y ← DoG → ←DoG→ SIFT(Lowe) Maxima and minima of the difference-of-Gaussian images detected by comparing a pixel (marked with X) to its 26 neighbors in 3x3 regions at the current and adjacent scales (marked with circles). SIFT-Keypoints Extraction Keypoints Matching Work Flow RANSAC Advantages/ Limitations Demo Transformation Examples SIFT Algorithm Introduction
  • 11. Gaussian pyramids Scale First Octave Scale Next Octave Difference of Gaussians Difference of Gaussian(DoG) SIFT-Keypoints Extraction Keypoints Matching Work Flow RANSAC Advantages/ Limitations Demo Transformation Examples SIFT Algorithm Introduction
  • 12. Keypoint Descriptors Feature Vectors  Thresholded image gradients sampled over:  16x16 array of locations in scale space  Histogram of 4x4 samples per window in 8 directions  Gaussian weighting around center  8 orientations x 4 x 4 histogram array = 128 dimensional feature vector  4x4 Gradient window SIFT Algorithm Keypoints Matching Work Flow RANSAC Advantages/ Limitations Demo Transformation Examples SIFT-Keypoints Extraction Introduction Gaussian Gradient Window
  • 13. Criteria for Selection  Prominent  Distinguishable-from neighborhood  Invariant  Stable to disturbances  Rare (exceptional) - from other selected points  Meaningful - with respect to image interpretation SIFT Algorithm Keypoints Matching Work Flow RANSAC Advantages/ Limitations Demo Transformation Examples SIFT-Keypoints Extraction Introduction
  • 14. SIFT keypoints  Detected Keypoints in reference and candidate image  Keypoints in both images will be matched Candidate image with keypointsReference image with keypoints SIFT Algorithm Keypoints Matching Work Flow RANSAC Advantages/ Limitations Demo Transformation Examples SIFT-Keypoints Extraction Introduction
  • 15. Keypoint Matching  Fundamental aspect in computer vision.  Based on Euclidean distance of the feature vectors.  Nearest neighbor algorithms.  Product between descriptors is calculated.  Inverse cosine of the products gives the Euclidean distance  Matches with Ratio of vector angles from the nearest to second nearest neighbor less than distRatio value are selected. SIFT Algorithm Keypoints Matching Work Flow RANSAC Advantages/ Limitations Demo Transformation Examples SIFT-Keypoints Extraction Introduction
  • 16. 90% of the false matches are removed. False matches due to ambiguos features or features arise from background clutter. Reliable object recognition with few best Matches Remove outliers Matched points SIFT Algorithm SIFT-Keypoints Extraction Work Flow RANSAC Advantages/ Limitations Demo Transformation Examples Keypoints Matching Introduction
  • 17. RANSAC: Algorithm RANdom SAmple Consensus  Estimate parameters of a mathematical model from a set of observed data which contains outliers.  A model is fitted using hypothetical inliers.  If other data fits to the model, added to the inliers.  Reestimated the model with the new set of inliers.  We used ransac-fit-homography which Robustly fits a homography to a set of matched points. SIFT Algorithm SIFT-Keypoints Extraction Work Flow Advantages/ Limitations Demo Transformation Examples Keypoints Matching RANSAC Introduction
  • 18. Ransac RANSAC: Result SIFT Algorithm SIFT-Keypoints Extraction Work Flow RANSAC Advantages/ Limitations Demo Transformation Examples Keypoints Matching Introduction
  • 19. Make the shortest image the same height as the other image. Append Images Transformation tform = cp2tform(M1,M2,’affine') SIFT Algorithm SIFT-Keypoints Extraction Keypoints Matching Work Flow Advantages/ Limitations Demo Transformation Examples RANSAC Introduction TStructure
  • 20. 2 3 Rod vs. OpenShrubs 1 Matches: 9 2 Rod: 626 keypoints found Image: 28958 keypoints found Matches: 16 …Good Matches SIFT Algorithm SIFT-Keypoints Extraction Keypoints Matching Work Flow RANSAC Advantages/ Limitations Demo Examples Transformation Introduction
  • 21. 2 Rod Vs Human Occlusion?? 1 Rod: 626 keypoints found Image: 2347 keypoints found Matches: 155 Rod: 626 keypoints found Image: 2347 keypoints found Matches: 144 2 3 …Good Matches SIFT Algorithm SIFT-Keypoints Extraction Keypoints Matching Work Flow RANSAC Advantages/ Limitations Demo Transformation Examples Introduction
  • 22. Limitations 1 Rod: 626 keypoints found Image: 17927 keypoints found Matches: 16 Matches: 4 2 Rod: 626 keypoints found Image: 30505 keypoints found 3 SIFT Algorithm SIFT-Keypoints Extraction Keypoints Matching Work Flow RANSAC Advantages/ Limitations Demo Transformation Examples Introduction
  • 23. Advantages  Locality: features are local, so robust to occlusion and clutter (no prior segmentation)  Distinctiveness: individual features can be matched to a large database of objects  Quantity: many features can be generated for even small objects  Efficiency: close to real-time performance  Extensibility: can easily be extended to wide range of differing feature types, with each adding robustness SIFT Algorithm SIFT-Keypoints Extraction Keypoints Matching Work Flow RANSAC Advantages/ Limitations Demo Transformation Examples Introduction
  • 24. Problems/Enhancement  Only invariant to affine transformations to a certain degree  Best performance on highly textured images  Use Principal components analysis PCA instead of Gaussian weighting for gradients Demo SIFT Algorithm SIFT-Keypoints Extraction Keypoints Matching Work Flow RANSAC Advantages/ Limitations Transformation Examples Introduction