CV Chapter 5 Local Features and Matching
Questions about the lecture 'Computer Vision' of the RWTH Aachen Chapter 5 Local Features and Matching
Questions about the lecture 'Computer Vision' of the RWTH Aachen Chapter 5 Local Features and Matching
Fichier Détails
| Cartes-fiches | 55 |
|---|---|
| Langue | English |
| Catégorie | Informatique |
| Niveau | Université |
| Crée / Actualisé | 04.02.2017 / 20.02.2017 |
| Lien de web |
https://card2brain.ch/box/20170204_cv_chapter_5_local_features_and_matching
|
| Intégrer |
<iframe src="https://card2brain.ch/box/20170204_cv_chapter_5_local_features_and_matching/embed" width="780" height="150" scrolling="no" frameborder="0"></iframe>
|
Créer ou copier des fichiers d'apprentissage
Avec un upgrade tu peux créer ou copier des fichiers d'apprentissage sans limite et utiliser de nombreuses fonctions supplémentaires.
Connecte-toi pour voir toutes les cartes.
What are two techniques to deal with outliers for recognition with local features? [2]
1. Random sample consensus (RANSAC) and 2. generalized hough transform
Define the RANSAC algorithm? [5]
1. Randomly select a seed group
2. Compute transformation from seed group
3. Find inlier to this transformation within threshold
4. Recompute least-squares estimate of transformation
5. Keep transformations with largest number of inlier
What are the probabilities for RANSAC? [2]
1. Probability that 1 sample of n points is correct: w^n with w fraction of inlier
2. Probability that k samples fail: (1-w^n)^k
What is a currently available application for indexing with local features?
Google goggles
What are the characteristic of Google goggles? [2]
1. Mobile visual search
2. Large-scale image matching problem
What is the parallel between indexing with local features and text retrieval? [2]
1. Find all pages on which a word occurs
2. Find all images in which a feature occurs
Name two solutions for indexing with local features? [2]
1. Inverted index files
2. Bag of words
What is the principle of inverted index files? [2]
1. Index features as point in high dimensional feature space
2. Search for close points in feature space
What is the complexity of search in feature space for inverted index files? [2]
1. Efficient for low-dimensional descriptors
2. Approximate for high-dimensional descriptors
How can the feature space be structured for inverted index files? [2]
1. Quantize feature space via clustering and defining prototype words
2. Vocabulary tree filled via training
What is the definition of term frequency - inverse document frequency (tf-idf) for inverted index files? [3]
1. Standard for text retrieval
2. Describe frame by frequencies of features within
3. t_i = n_id/n_d *log(N/n_i)
n_i number of occurrences of i in database
n_id number of occurrences of i in document d
n_d number of occurrences in document d
N total number of documents in database
What is the principle of bag of words? [3]
1. Split several images into visual words removing spatial layout
2. For every image specify a histogram of all words // Vector
3. Compare histograms with any known measures
To which technique is bag of words similar to?
Similar to bag of textons for texture representation
What can be done to save the spatial layout in bag of words? [2]
1. Spatial pyramid representation
2. Perform BoW for each sub-gird of an image
How is the performance of bag of words compared to the parts and shape model?
Better
What are the motivations for local feature representations? [2]
1. Global representations have limitations
2. Local representations have increased robustness to occlusion, articulation and intra-category variation
Name two applications of local feature representations? [2]
1. Analyze if two images display same content from different point of view
2. Image stitching
What are the steps of image stitching? [3]
1. Detect feature points
2. Find corresponding pairs matching local descriptors
3. Align images
What are the steps of feature point detection (step one) of image stitching? [3]
1. Define region around keypoint
2. Extract and normalize region content
3. Compute local descriptor
What are the steps of pair matching (step two) of image stitching?
Find same points // Correctly find corresponding one
What are local feature detectors in recent applications in computer vision?
Basic building block
What are the requirements for local feature detectors? [5]
1. Regions are repeatable and accurate
2. Invariant to translation, rotation and scale changes
3. Features have locality making them robust to occlusion and clutter
4. Provide sufficient number of distinct regions
5. Operations must be close to real-time performance
Name available local feature detectors? [7]
1. Hessian and Harris
2. Laplacian, DoG
3. Hessian-/Harris-Laplace
4. Hessian-/Harris-Affine
5. EBR and IBR
6. MSER
7. Salient Regions
What are the goals of keypoint localization for local feature detectors? [3]
1. Repeatable detection
2. Precise localization
3. Interesting content
What are the basic techniques for keypoint localization for local feature detectors? [3]
1. Look for 2D signal changes
2. Around corner two or more dominant gradients exist
3. Slide small window
What are the three cases in keypoint localization for local feature detectors? [3]
1. Flat with no changes in all directions
2. Edge with no changes along edge direction
3. Corner with significant change in all directions
What are the definitions of the intensity shift change E(u,v) for the harris detector? [7]
1. E(u,v) = Sum_x,y w(x,y)[I(x+u,y+v)-I(x,y)]²
2. E(u,v) = [u,v]*M*[u;v]
With
3. w(x,y) is window function
4. I(a,b) is intensity function
5. M = Sum_x,y w(x,y)*[Ix²,IxIy;IxIy,Iy²]
6. M = [SumIx²,SumIxIy;SumIxIy,SumIy²]
7. M = Sum [Ix;Iy]*[Ix,Iy]
What holds for the window function for the harris detector? [3]
1. Not rotation invariant if rectangular
2. Rotation invariant with Gaussian g(sigma)
3. Not scale invariant
What is the definition for an axis-aligned corner for the harris detector? [2]
1. M = [l1,0;0,l2]
2. If one of l1 or l2 is close to zero → no corner
What is the definition for an general corner for the harris detector? [4]
1. M = R^-1*[l1,0;0,l2]*R // Eigenvalue decomposition
2. R = det(M)-a*trace(M)² = l1*l2 – a(l1+l2)²
3. Faster with no eigenvalues and a\in[0.04,0.06]
4. Visualization of M as ellipse
-
- 1 / 55
-
