Homework: Cameras and Convolution Solution






  • This is an individual assignment. ’Individual’ means each student must hand in their own answers, and each student must write their own code in the homework. It is admissible for students to collabo-rate in solving problems. To help you actually learn the material, what you write down must be your own work, not copied from any other individual. You must also list the names of students (maximum two) you collaborated with.


  • You must submit your solutions online on Canvas. We recommend that you use LATEX, but we will accept scanned solutions as well. Please place your homework (.pdf), code, images and any additional files into the top level of a single folder named <PennKey>.zip


  • Notation Clarification For notations in all questions below, we denote I as input image, f as ker-nel and g as output image. In addition, questions are independent of each other except additional notifications.


  • Start early! If you get stuck, please post your questions on Piazza or come to office hours!


1     Getting Started


1.1   Introduction to MATLAB / Python


In this question, you will be required to build upon the starter code provided, to read in an image and apply the Sobel operator to it. The kernel which highlights the vertical edges when convolved with an image has been given to you in the code. You need to perform the following tasks:


  1. Read in the image Bikesgray. j pg into the variable img1


  1. Convolve the image with the given kernel f 1


  1. Display and save the result of the convolution


  1. Come up with a kernel f 2 similar to f 1, but one which causes the horizontal edges to be highlighted


  1. Convolve the image img1 with kernel f 2


  1. Display and save the result of the convolution.


  • Question 1.1: Please implement the above via MATLAB or Python. Submit your code and the two images generated as a result of the convolution with kernels f 1 and f 2 respectively.


1.2   Filtering


Recall the definition of filtering is as follows:


g(i, j) = ÂI(i + m, j + n) ⇤ f (m, n) = I ” f (1)



where I is the image, f is the kernel for filtering.



  • Question 1.2: Using the I and f given below, check if the commutative property (I ” f ⌘ f ” I) holds for the filtering operation. Show by hand how you arrived at your answer. Assume zero-padding along the boundary, and ’same’ output size.


I = 2 0.5 1.0 0.0 3 (2)
0.5 2.0 1.5
f = 42.0 0.5 1.05 (3)
2 0.0 1.0 0.5 3
0.5 1.0 0.0
40.5  0.0 0.55  

Note: The matrices for I and f given in Eq. 2 and 3 respectively are to be used only for Question 1.2.


2     Convolution


Recall the definition of convolution,


g = I ⌦ f (4)

where I and f represents the image and kernel respectively.


Typically, when kernel f is a 1-D vector, we get
g(i) = ÂI(i − m) f (m) (5)
where i is the index in the row or column dimension.
If the kernel f is a 2-D kernel, we have
g(i, j) = ÂI(i − m, j − n) f (m, n) (6)


where i and j are the row and column indices respectively.


In this section, you need to perform the convolution by hand, get familiar with convolution in both 1-D and 2-D as well as its corresponding properties.


Note: All convolution operations in this section follow except additional notifications: 1. Zero-Padding, 2.


Same Output Size, 3. An addition or multiplication with 0 will count as one operation.



For this problem, we will use the following 3 ⇥ 3 image: 0.0 3
I = 22.0  1.0
0.0 1.0 1.0
40.0 3.0 −1.05
You are given two 1-D vectors for convolution: T
fy = 1.0  1.0  1.0
fx = −1.0  0.0  1.0

Let g1 = I ⌦ fx ⌦ fy, fxy = fx ⌦ fy and g2 = I ⌦ fxy.


Note : fxy should be full output size.











  • Question 2.1: Compute g1 and g2 (At least show two steps for each convolution operation and intermediate results), and verify the associative property of convolution
  • Question 2.2: How many operations are required for computing g1 and g2 respectively? Show addition and multiplication times in your result.


  • Question 2.3: What does convolution do to this image?


3   Kernel Estimation


Recall the special case of convolution discussed in class: The Impulse function. Using an impulse function, it is possible to ’shift’ (and sometimes also ’scale’) an image in a particular direction.


For example, when the following image 2 a b c 3
I = d e f (10)
is convolved with the kernel, 4 g h i 5
2 1 0 0
f = 0 0 03 (11)
it results in the output: 4 0 0 05  
2 e f 0
g = h i 03 (12)
4 0 0 05  


  • Question 3: Using the two tricks listed above, estimate the kernel f by hand which when convolved


with an image 2 1 5 2 3
I = 7 8 6 (14)
results in the output image 2 4 3 9 4 5 3
29 33 10
g = 62 52 30 (15)
4 15 45 20 5
IQ = 21  1 1 0 3, IT = 20  1 1 13
  6 1 1 1 0 7 6 0 0 0 0 7
0 0 0 0 0 1 1 1
6 7 6 7
4 1 1 1 0 5 4 0 1 1 1 5


Instinctively, the human eye can automatically detect a potential barrel in the top left corner of the query image but a computer can’t do that right away. Basically, if the computer finds that the difference between

query image’s features and the template’s features are minute, it will prompt with high confidence: ’Aha! I have found a barrel in the image’. However, in our circumstance, if we directly compute the pixel wise distance D between IQ and IT where


D(IQ, IT ) = Â(IQ(i, j) − IT (i, j))2 (16)
i, j


we get D = 10 which implies that there’s a huge difference between the query image and our template. To fix this problem, we can utilize the power of the convolution. Let’s define the ’mean shape’ image IM which is the blurred version of IQ and IT .

IM = 200.5 1 1 0.5 3
6 .25  0.5  0.5 0.25 7
00.25  0.5  0.5 0.25
6 7
4 .5 1 1 0.5 5
• Question 4.1: Compute two 3 ⇥ 3 convolution kernels f1, f2 by hand such that IQ ⌦ f1 = IM and

IT ⌦ f2 = IM where ⌦ denotes the convolution operation. (Assume zero-padding)


  • Question 4.2: For a convolution kernel f = ( f1 + f2)/2, we define IQ0 = IQ ⌦ f and IT0 = IT ⌦ f . Compute IQ0, IT0 and D(IQ0, IT0 ) by hand. Compare it with D(IQ, IT ) and briefly explain what you find.


  • Camera Model and Camera Projection


  • Question 5.1 Camera Sensor Size: Choose a digital camera that you own, for example: a mobile phone. What is the height (mm) and width (mm) of its image sensors (front-facing and back)? You can find this by looking it up on the internet. Please include the web address as a reference. How does the sensor size affect the field of view of the camera?


  • Question 5.2 Pixel Size: What is the image sensor resolution of the back camera? Compute the size of a pixel in millimeters.


  • Question 5.3 Focal Length: What is the focal length of your camera (front-facing and back)? You can find this by looking it up on the internet. Please include the web address as a reference. Also, compute the focal length of your back camera by measuring its field of view and using the size of its sensor. How similar are they? Ignore the difference caused due to autofocus. How does the focal length of the camera affect its field of view?


  • Question 5.4 Camera Matrix: Compose the intrinsic camera matrix for your digital camera. As-sume the axis skew to be zero.


  • Question 5.5 Depth of an Object: Using your digital camera, capture a picture of your friend (say F) with known height (in meters) standing some distance in front of Benjamin Franklin (say B) in front of College Hall at the University of Pennsylvania as shown in the Figure 5.5(a) and Figure 5.5(b). Please ensure that the camera plane is perpendicular to the ground plane while capturing the image.



  • Question 5.5.1: Given the height HF (in meters) of F and pixel height hF , compute the distance from F to the camera (in meters). Show all your work.


  • Question 5.5.2: Given the height HB (213 inches) of B and pixel height hB, compute the distance from B to the camera (in meters). Show all your work.


  • Question 5.6 Dolly Zoom: Capture two images with different camera positions in the same setting as Question 5.4, taking a few steps, Dd backwards for the second image as shown in Figure 5.6(a). F and B should appear smaller in the second image since Dd is positive. Simulate the Dolly Zoom effect by scaling up and cropping the second image as shown in Figure 5.6(b) such that your friend is the same pixel height in both the images.




Figure 5.6(b): Scaling up and Cropping


Figure 5.6(a): Geometry of Dolly Zoom


  • Question 5.6.1: Given pixel height h0B in the first image, compute the pixel height of B, h00B in the second image. Show all your work.


  • Question 5.6.2: Measure the pixel height of B, h00B in the second image. Are they close to each other?
  • Question 5.6.3: If we want to increase h0B three times while keeping h0F the same, what should be the new camera position and focal length of the camera? Assume that your camera has an optical zoom.


  • Question 5.6.4: (Optional) Create a .gif of the Dolly Zoom effect in the given setting by cap-turing more than two images.

error: Content is protected !!