Visual computing part 1

What’s the location of pixel (x,y) in memory using rgb?

3 x (y x w + x)

What’s the location of pixel (x,y) in memory using greyscale?

(y x w + x)

What do HSL and HSV stand for?

Hue, saturation, lightness/value.

Fourier transform linearity

Fourier transform shifting

g(x) = f(x-a)

Fourier transform modulation

g(x) = e^{iax} f(x)

Fourier transform scaling

g(x) = f(ax)

Fourier transform differentiation

do (iw)^n for higher order derivatives

Fourier transform convolution

What is the relation between focus(f), object distance(u), and image distance(v)

What are the two types of image sensor?

CCD and CMOS

How do LCD displays work?

Twist polarisation, max energy = 90 degress, min light

How do we workout total number of colour for an 8-bit image?

How do we distribute quantisation error over multiple pixels?

Error diffusion!

What is a horizontal gradient sobel filter?

What is a vertical gradient sobel filter?

What is a diagonal gradient sobel filter?

How does erosion work?

Only keep a pixel white if all surrounding pixels are white.

What is opening?

erosion then dilation

What is closing?

dilation then erosion

What is euler’s formula?

What is the Fourier transform useful for?

• Low and high pass filtering

• Linear filtering

• Removing structured noise

• Compression

What is the complexity of the fourier transform?

naive O(n^2), can split into two inputs of size n/2 which gives us O(n \\log n)

What is uniform scaling?

Scalar is the same for each component (axis)

Formula for anti-clockwise rotation?

Formula for clockwise rotation?

Formula for shear/skew

What are projective transformations? How do we do them?

1. Parralel lines are not preserved

2. Fill in the bottom row of transformation matrix

What is the formula for a 3d rotation about x axis?

What is the formula for a 3d rotation about y axis?

What is the formula for a 3d rotation about z axis?

How do we project a 3d transformation through a pinhole onto the image plane?

1. Apply the transformation based of the focal length, and then transform so the result is in the top right, this is called the intrinsic camera matrix.

2. We then add a 3d transform from world to camera-co-ordinates.

What is the difference between perspective and orthographic projection?

Perspective parralel lines don’t stay parallel, orthographic they do

What are the different ways we can represent 3d projections?

• 3x3 matrix, not a minimal representation

• Euler angles, minimal parametisation, has gimbal lock

• Axis- angle, no gimbal lock minimal representation

What is openGL?

is cross-language/platform and is most widely used and supported

What is webGL?

is a JavaScript API this is based on OpenGL but can run in browser and dosen’t need to be compiled

What is three.js?

is an easier to use wrapper for webGL.