HW4 - CS 556: HOMEWORK 4 due March 15, 2010 This homework...

Info iconThis preview shows pages 1–2. Sign up to view the full content.

View Full Document Right Arrow Icon
CS 556: HOMEWORK 4 due March 15, 2010 This homework is about object recognition using the bag-of-words model, on the Graz-02 im- ages of bikes, cars, and people, which can be downloaded from: http://lear.inrialpes.fr/people/marszalek/data/ig02/ Images of the three object classes are divided into training and test sets. Use the training images to learn the object representation, and use the test images to evaluate your representations. 1. Extract the HOG features of a regular grid of patches from each image. 2. Use K-means clustering to find a dictionary of 300 visual words based on the HOG features extracted from all training images of all object classes. 3. Learning 1: Detect the codewords in the training images, and estimate the histograms, H 1 ,H 2 ,H 3 , of occurrence of the 300 codewords per each of the three object classes. 4. Learning 2: Partition each training image into non-overlapping windows using the quad-tree of three levels, resulting in a total of M = 1 + 4 + 16 = 21 windows. Then, treat each
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Image of page 2
This is the end of the preview. Sign up to access the rest of the document.

Page1 / 2

HW4 - CS 556: HOMEWORK 4 due March 15, 2010 This homework...

This preview shows document pages 1 - 2. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online