Questions tagged [sift]
The sift tag has no usage guidance.
58
questions
0
votes
0
answers
38
views
How to Implement the Sub Pixel Refinement Phase in Scale Invariant Feature Transform (SIFT)
I'm trying implement SIFT(Scale-invariant feature transform) myself. I've read some documents and stuck at Sub-Pixel Refinement - left without knowing how to actually implement it in code. I do ...
3
votes
1
answer
116
views
The SIFT Descriptor and Image Resolution
When reading about SIFT I read halving resolution is the same as increasing $\sigma$ of the gaussian in terms of feature detection but reducing resolution has the advantage of reducing processing.
...
- 818
6
votes
1
answer
452
views
Why is scaling of images / pixels into `[0, 1]` range performed before SIFT (Scale Invariant Feature Transform) algorithm?
The SIFT paper and the paper of Anatomy of the SIFT Method do not mention that the input images should be preprocessed (normalized, re-scaled) before feeding images into the standard SIFT algorithm. ...
- 213
3
votes
1
answer
186
views
Why resizing an image smoothed by Gaussian by factor of 2 also increase sigma by factor of 2
In the paper about SIFT algorithm (Distinctive Image Features
from Scale-Invariant Keypoints) it says at the end of section 3.3:
We double the size of the input image using linear interpolation prior ...
- 31
2
votes
1
answer
192
views
What's the difference between SIFT and general stereo matching algorithm (eg, sgbm)?
I am working on an open source project s2p which creates digital height model from satellite stereo imagery. The procedure of how s2p works can be roughly summarized with the following steps.
Split a ...
- 213
2
votes
1
answer
207
views
Registration of point cloud based on feature matching method
I recently find myself an interesting problem to solve.
Basically I got 2 partially overlapped point clouds (in real world) that I want automate the registration process. My naive approach currently ...
- 21
1
vote
0
answers
77
views
Are there algorithms that can match SIFT's key points of 2 images using their mutual arrangement?
I'm solving image matching task using SIFT as a feature extractor
To match obtained descriptors I used FLANN based Matcher (provided by OpenCV) and result looks good:
But applying same algorithm to ...
1
vote
0
answers
32
views
feature extraction techniques for iris recognition
I want to ask how I can divide feature extraction techniques to feature detectors and feature descriptors. I have big problem how to understand it. For example I can use Gabor filters (feature ...
- 73
3
votes
1
answer
674
views
Why do we need to construct Gaussian pyramid using SIFT detector
I am learning about SIFT detection and descriptor. I am slightly unsure about why a Gaussian pyramid is built for the image.
I do understand that within each octave, we are applying the Difference of ...
- 283
1
vote
1
answer
159
views
Invariances of FFT-based Image-Registration vs. SIFT-Features
TL;DR: I don't understand how invariant FFT-based image-registration techniques are to object alterations (scratches, marks etc.) in comparison to SIFT-features.
I want to build kind of a feature-...
- 201
3
votes
2
answers
969
views
Why do we need multiple layers in each octave and multiple octaves in SIFT?
I skimmed through the SIFT paper. I understand that there are multiple octaves, which are composed of multiple layers. The layer $k$ of an octave (btw, where does this name come from?) corresponds to ...
user40095
0
votes
0
answers
85
views
SIFT between 1 image and dataset (2 images)
I use SIFT, and I want to find the angle of inclination of an image compared to two reference images, in short:
I have 3 images :
Image x: input image
Two images 1 and 2, reference images that are ...
- 11
0
votes
1
answer
275
views
5
votes
1
answer
1k
views
Laplacian of Gaussian Approximation and Gaussian Blur as the Solution of Heat Equation
While I was reading SIFT paper(Lowe, 2004), I came across the method that he apply "heat diffusion equation" to Gaussian function to derive that
$$ \frac{∂G}{∂σ}
= σ∇^2G
$$
I searched Wikipedia and ...
- 153
1
vote
1
answer
310
views
How to obtain the relative position (coordinates) of SIFT feature from image center coordinates?
I have extracted SIFT features (keypoints and descriptors) of training set and saved in an inverted index form.
My question is how can I calculate the coordinates of each visual features in an image ...
- 161
1
vote
0
answers
271
views
VLFEat: How to extract the SIFT descriptor without SIFT feature detection?
I am using the MATLAB interface to the VLFeat toolbox in order to calculate a SIFT descriptor for image patches extracted around key points. Due to the greater framework of what I'm doing, I cannot ...
- 233
2
votes
2
answers
2k
views
Why is scale space (DoG) needed to detect scale invariant features?
I would have a theoretical question, about the detector part of the SIFT algorithm, which as it is explained by D. Lowe in his paper, used the DoG to detect keypoints in an input image.
Problem ...
- 21
1
vote
1
answer
420
views
How to recognize an object from a small training set of images?
Given a small training set of images (say, around 6 max) of the same object, how to measure how likely is another query image to contain the given object?
The training images are "clean", i.e. only ...
- 111
2
votes
3
answers
2k
views
Intuitive understanding of scale-space extrema detection
Can someone explain intuitively why local maxima and minima in the scale-space domain make for good keypoints?
I understand using LoG or DoG zero-crossing points to identify spacial variations, i.e. ...
- 206
1
vote
0
answers
54
views
Analytical expression for salient Blob detection in scale space
I am working on exact mathematical expression which can be obtained in image processing tasks.(I do not know much about image processing but my work is more mathmatical) In Gaussian scale space, we ...
- 111
0
votes
1
answer
248
views
Hessian Matrix. Second partial derivative test
I have a question about the hessian matrix / determinant of the hessian matrix. For the Hessian matrix, one is able to determine the local curvature around a point. And with the Second partial ...
- 55
0
votes
0
answers
46
views
Counting percentage of fouling on a surface
I am completely new to image processing and python. I hope I can find some help here. I am trying out a project on detecting fouling% on a ship hull using raspberry pi. I am trying to learn how this ...
- 13
6
votes
1
answer
115
views
What Is "Description Vector" in Image Processing?
When we want to use classifiers like SVM, we first should extract descriptors using algorithms like SIFT.
But I have a question which might call a silly one:
Let's assume
$$
\begin {equation}
D_a=\{...
- 711
4
votes
2
answers
5k
views
Why optical flow ? why not tracking
I have a question that might be stupide !
Supposing that I try to detect object moving in a video or human action.
Many works are based on optical flow computation.
My question is why using OF is ...
1
vote
1
answer
159
views
Image Projection with SIFT
Recently, I am thinking about a problem regarding image projection. I have a larger picture as bellow.
I want to find the transformation matrix, x'=Hx. When the above image pass the transformation ...
- 131
3
votes
2
answers
275
views
Would the locations of SIFT features generally agree with features detected by Shi-Tomasi method?
I am having a comparison of different feature detection methods. So if we compare the SIFT (Scale-invariant feature transform) and Shi-Tomasi method, will their feature locations agree and why? Are ...
- 131
2
votes
1
answer
2k
views
Using ROC curves for comparing the performance of SIFT and SURF
I want to compare the performance of SIFT and SURF against some standard image set.
After a bit of research, I found that ROC curves are one of the best ways to do this task.
In order to plot ROC ...
- 23
0
votes
1
answer
304
views
Is there two Types of Scale Space in SURF algorithm?
I couldn't figure out below paragraph on SURF paper and hope that someone can help me to understand it.
Bay H., Ess A., Tuytelaars T. Van Gool L. - Speed-Up Robust Features
(SURF), page 4, ...
- 151
2
votes
1
answer
240
views
Can SIFT run in realtime?
What is the best possible runtime for extracting SIFT keypoints and Descriptors?
I know it depends on #keypoints extracted. So, say for image of size 640x640, the code that I have been using requires ...
- 127
4
votes
1
answer
197
views
Why Does the Odd Multiple of $ \frac{\pi}{4} $ on Gaussian Cause Loss in Repeatability Under Image Rotations?
I couldn't figure out below paragraph on SURF paper and hope that someone can help me to understand it. Why image rotations around odd multiples of $\frac{\pi}{4}$ lead to a loss of repeatability?
...
- 151
0
votes
1
answer
79
views
need general resource about image matching
i'm working on image matching mainly SIFT and SURF for my thesis. i read Moravec's and harris corner detectors and understand them quiet well in my own opinion.A COMBINED CORNER AND EDGE DETECTOR[1]
...
- 151
0
votes
1
answer
1k
views
k-NN match for object recognition
The $\frac{d_1}{d_2}<0.6$ gives a large number of false matches and it assumes the nearest neighbour is a correct match.
Anybody knows any good method to reduce number of false matches or filter ...
- 127
0
votes
1
answer
499
views
Which is the best method for comparing the histogram obtained after BOW?
Euclidian distance
Cosine similarity
Histogram Intersection
Or any other?
Consider the importance of the method in case of images. Eg, Cosine Similarity matches angle of two vectors...
What impact ...
- 127
1
vote
2
answers
1k
views
what is the output of BoW after an image has been trained with SIFT algorithm and k-means
SIFT algorithm provides a 128 dimensional feature vector that is used for image classification.When all the interest points(key points) are taken together and K-means clustering is applied,the image ...
- 109
1
vote
1
answer
1k
views
What steps come after finding the gradients of the SIFT key points?
I am using a SIFT algorithm to extract features from an image. I understand that the SIFT descriptor first finds the extrema and then finds the gradient and direction of each of these interest points....
- 109
0
votes
0
answers
656
views
Distance-metric for SIFT descriptors matching
Will using Cosine-Similarity or Histogram Intersection as distance-metric, for matching SIFT descriptors, help to get better results?
Anybody tried on that ever?
- 127
0
votes
2
answers
226
views
Increasing Recall rate for SIFT
Is there a better way to increase Recall Rate when using SIFT features?
I am thinking a way to replace the NN1/NN2 ratio to account for slightly distorted objects.
Moving towards clustering and using ...
- 127
2
votes
3
answers
1k
views
Visualize visual words in bag of words model
I am implementing visual bag of words through these steps:
Find interest points using SIFT
Calculate SIFT descriptor
Build codebook through kmeans clustering of SIFT descriptors.
How can I visualize ...
- 223
0
votes
1
answer
584
views
Segmentation step when using feature descriptors
In my final year project, we are using a SIFT feature descriptor to classify and recognize objects, but I have never come across segmentation in the process of image classification/categorization.
...
- 109
2
votes
1
answer
6k
views
SIFT - Taylor Expansion
I'm trying to implement SIFT algorithm.
After I found the min/max points in DOG images I need to find the real min/max subpixels by taylor exapnsion:
$$
\begin{equation}
D(\mathbf{x}) = D + \frac{\...
- 23
3
votes
0
answers
3k
views
Creating a Gaussian Pyramid
I'm trying to recreate a gassuain pyramid using the following scales:
There seems to be two concepts used here:
1) When an image is halved, applying a gaussian kernel of σ will apply as 2σ.
2) When ...
- 153
9
votes
1
answer
391
views
What is the story behind the story about SIFT descriptor?
The following is from Lowe 2004 paper ( http://www.cs.ubc.ca/~lowe/papers/ijcv04.pdf ).
One obvious approach would be to sample the local image intensities around the keypoint at the appropriate ...
- 457
6
votes
2
answers
2k
views
SIFT - why s+3 scales per octave?
I have a problem with SIFT that I do not understand. Lowe [1] proposed in his work the s=3 levels of scale are enough for one octave. Afterwards, he mentioned that ...
- 113
4
votes
2
answers
6k
views
Why is it necessary to implement octaves in sift
I've been studying sift very hard for two weeks.
I found much materials about scale space.
It is very hard to understand scale space depthly.
What I've found and confusing things are that
'scale' in ...
- 457
1
vote
1
answer
183
views
I have a question on SIFT
I am reading a paper on SIFT by Lowe.
In it, gradient magnitudes and orientations are calculated using these formulae:
$$$$
I don't know why m(x,y) is calculated above. I heard that gradient ...
- 457
0
votes
1
answer
2k
views
taylor expansion of scale space function
I see the following expression from http://en.wikipedia.org/wiki/Scale-invariant_feature_transform
The quadratic Taylor expansion of the Difference-of-Gaussian scale-space function, with the ...
- 457
1
vote
1
answer
447
views
Keypoint position and scale in SIFT
If I understood it right, a keypoint is a tuple $$(x,y,\sigma, r),$$ where $x,y$ define the position of the keypoint and $r$ an orientation (given by the most domiant gradiant vector around the ...
- 11
2
votes
0
answers
185
views
Affine detector + SIFT descriptor
I am approaching to the Oxford Buildings dataset which contains already computed SIFT descriptors from affine detected keypoint locations.
Usually SIFT detector provides to its detected keypoints ...
- 151
10
votes
3
answers
3k
views
Detecting a filled glass object
I've been sent here from this question in stackoverflow, please excuse me if the question comes too specific and it's not in the manners in here:)
The task is to find a glass with specific liquid in ...
- 202
1
vote
3
answers
1k
views
How to proceed for BOW creation From sift features
I extracted sift feature to construct a bag-of-words model. I did the following to extract the feature vector:
Extracted features from the key frames.
After the feature extraction I tried using ...