The document discusses two algorithms for object detection: HOG and SIFT. HOG (Histogram of Oriented Gradients) focuses on the shape of an object by using the magnitude and direction of gradients to generate histograms and compute features. SIFT (Scale Invariant Feature Transform) describes local image areas by extracting invariant features to generate a set of key points for matching objects across different scales and rotations. Both algorithms can be used to detect objects by matching image features to trained models.

1. OBJECT DETECTION
HOG AND SIFT ALGORITHM

2. HOG ALGORITHM
 HOG stands for histogram of oriented gradients.
 The hog descriptor focuses on structure or shape of the object.
 It uses magnitude as well as direction of the gradient to compute the features.
 It generates histogram by using magnitude and direction of the gradient.

4.  Here we calculating gradient magnitude and direction, to calculate pixels intensity we need
 X direction=|40-70|=30
 Y direction=|20-70|=50
 By these values we are calculating magnitude and direction of the gradient
 By using magnitude and direction we calculate feature vectors
20
40 70
70

7. e

8.  Before getting the hog feature and after concatenating feature vectors we are supposed to do
normalize.
 Suppose we have taken 150*300 pixels and multiply with 2 to increase the brightness and divided
by 2 to decrease the brightness, then you cant compare two images without normalization bec’z
the pixels intensity will be changed.
 But if you normalize the feature vectors it is easy to compare

10. • For hog features giving human template and giving output for convolving with human model
Then it will predict whether it is human or not.

11. SIFT ALGORITHM
 SIFT stands for scalar invariant feature transform and was first presented in 2004, by D.Lowe,
University of British Columbia.
 It is a way to describe a local area in an image.
 In this whole image is reduced to set of points.
 SIFT is invariance to image scale and rotation.
 This algorithm is patented, so this algorithm is included in the Non-free module in OpenCV.

12.  Major advantages of SIFT are
 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 a wide range of different feature types, with each adding
robustness.

13.  The scale space of an image is a function L(x,y,σ) that is produced from the convolution of a Gaussian
kernel(Blurring) at different scales with the input image.
 Within an octave, images are progressively blurred using the Gaussian Blur operator.
 Mathematically, “blurring” is referred to as the convolution of the Gaussian operator
and the image.
 Gaussian blur has a particular expression or “operator” that is applied to each pixel.
What results is the blurred image.
gaussian blur
Gaussian blur operator

14.  In difference of gaussian kernel(DOG) we use those blurred images to generate another set of
images
 These dog images are used to find interesting points in the image.
 These process is done for all images in the gausian pyramid.
 After that we stack those different images on top of the each other and basically your looking
extreme points.
 In these locally distinct which stand out and those are your key points.