Sidebar

- Harris interest points - definition, algorithm for detection, parameters. Explain the motivation behind the definition. Describe the effects of the parameters on the number of detected points. To which transformation (geometric/photometric) is this detector invariant?
- Describe the algorithm for selection of interest point (region) scale using the Laplacian.
- Describe steps to generalize Harris detector to become affine invariant.
- Hessian and Difference of Gaussian interest points. Definition, properties.
- Define Maximally Stable Extremal Regions (MSER). Describe the algorithm for their detection. Properties of extremal regions end the maximally stable subset.
- The FAST interest point detector
- The SIFT descriptor. Describe the algorithm and its properties.
- Describe “Local Binary Patterns” like descriptors.
- How are local affine frames used for invariant description?
- Wide-baseline matching. Describe the steps for obtaining correspondences between a pair of images, which are taken from different viewpoints.
- How to find similar descriptors in sub-linear time?
- How does the “bag-of-words” method work?
- What is the “inverted file” and how it is used for the image retrieval?
- Define the tf-idf reweighting for visual words.
- Describe the “query expansion” mechanism for improving the recall of the image retrieval.
- Describe how the min-Hash method describes the images. Which properties it has?
- Describe the RANSAC algorithm, its properties, advantages and disadvantages. Which parameters it has?
- Describe the steps for object detection using “sliding windows” (“scanning windows”). How is the reasonable speed achieved?
- Describe how to use an integral image for computing the sum of the intensity and the intensity variance for a rectangular region.
- Why is the Adaboost algorithm often used for the “sliding window” methods? Give more than one reason.
- Describe the Hough transformation algorithm for detection or parametrized structure (line, circle, …). Discuss the properties of the algorithm (time and memory requirements, parameters).
- Compare the Hough transformation with a brute-force search algorithm.
- Compare the Hough transformation with RANSAC.
- For a static scene and viewing by camera with only horizontal movement. Draw a image patch, which will be useful for a tracking using a gradient method (KLT tracker). Which properties should has such image patch to be suitable for tracking?
- Which image patches are suitable for tracking? Why? Which patches are not suitable?
- Mean-shift algorithm. Describe the principles and simulate calculation for 1D example.
- Mean-shift algorithm. Color pixels [R,G,B] represented in 3D space. How you can reduce the color-space into 256 color-space?
- DCT - discriminative (kernel) correlation tracking. The algorithm, representation of the object, the search method.
- DCT tracking in the presence of rotation and scale change.
- Deep Neural Nets for image classification. Structure - convolutional, pooling and fully connected layers. Non-linearities.
- Deep Neural Nets for image classification. Learning - the cost function, the SGD (stochastic gradient method), drop-out, batch normalizaton. SGD parameters.
- Deep Neural Nets for detection. Proposal-based and end-to-end methods. Class label and bounding box prediction.
- Deep Neural Nets - applications in computer vision.

courses/mpv/labs/exam_questions.txt · Last modified: 2018/02/19 15:05 (external edit)