133 team check 1

Why might HSV be preferred over RGB for certain computer vision tasks?

HSV separates (hue) from (value) and (saturation), making it easier for algorithms to distinguish colors, while RGB considers them all together.

Describe hue, saturation, and value in your own words.

Hue represents the color itself, saturation indicates the intensity or purity of the color, and value reflects the brightness or lightness of the color.

Why must RGB values be normalized to the [0-1] range in the rgb_to_hsv function?

HSV operates on a scale of [0,1] so the saturation and value outputs are valid.

What are max, min, and delta, and why are they important?

max is the highest value, min is the lowest value, and delta is the difference between max and min for values of R,G, and B. ; they are important for calculating hue, saturation, and value in the HSV color model.

Why should the rgb_to_hsv function return floating-point values instead of unit8?

unit 8 can only store whole numbers, and using it would lose / round the decimal values needed from 0-1

How are you applying the rgb_to_hsv function to every pixel?

Your are doing this through the convert_to_hsv function using a nested for loop, and then calling that function in the main function.

Why is resizing and padding to a standard size necessary?

Resizing and padding ensures all images have the same dimensions, while not distorting them

What would happen if the input image to the convert_to_hsv were grayscale?

It will cause an error in the program since it only has two dimensions, compared to a 3 channel color image (check)

How do you construct a normalized mean kernel?

Create a matrix of the number 1 with the given dimensions, then divide each number by the total pixel number

How does a kernel perform filtering on an image?

it slides across every pixel in an image, multiplies its numbers by the image numbers and then adds them to make up the final picture

What is an image kernel?

A small matrix used to give visual effects to an image

What does it mean to use a sliding-window operation?

It means moving a small window over a larger image, evaluating one area at a time.

How does padding affect output image size?

Padding either keeps the image size the same or increases it

How do you determine the amount of zero-padding required?

Divide the kernel size by 2 and round the answer down, which determines the amount of pixels needed at the border of the image

What is stride, and what does stride = 1 mean?

Stride is how many pixels the window moves at a time, a stride of one means the window shifts by one pixel for each calculation.

Why must kernel_size be an odd number?

It must be odd so there is a center to the matrix that fits a pixel clearly

Explain the difference between mean blur and Gaussian blur in terms of weighting.

Mean blur gives equal value to surrounding pixels, while gaussian filtering gives different weight to surrounding pixels depending on their closeness

Why is blurring typically performed before feature extraction steps like edge detection?

blurring smoothes out any noise, making sure the images edges are accurate

Why is Gaussian blur typically preferred over mean blur for this workflow?

A mean blur would create blocky, imperfect image edges, while a gaussian blur can smooth out these imperfections and make the edges look more sharp.

What effect does sigma have on the Gaussian kernel?

sigma controls the distribution of the kernel, which affects the strength of the blur