How many raw pixel values?

Background

An image is a dense grid of numbers. The total number of raw values is the product of its dimensions:

$$\text{values} = C \times H \times W$$

A grayscale image has $C = 1$; an RGB image has $C = 3$. This count is what you'd get if you flattened the image into a single feature vector.

Problem statement

Implement raw_value_count(H, W, C) returning the number of scalar values in the image.

Input

• H — int, height.
• W — int, width.
• C — int, number of channels (1 grayscale, 3 RGB).

Output

Returns an int: $C \times H \times W$.

Examples

Example 1 — an RGB ImageNet crop

Input:  H = 224, W = 224, C = 3
Output: 150528

Explanation: $3\times224\times224 = 150{,}528$.

Example 2 — a grayscale MNIST digit

Input:  H = 28, W = 28, C = 1
Output: 784

Constraints

• Return H * W * C as a plain int.

Notes

• 150,528 inputs is exactly why feeding raw images to a dense MLP explodes the parameter count — and why CNNs, which share weights, are used instead.