1D Convolution (Valid)
HardMachine LearningArraySliding Window
Description
Given a 1D signal and a 1D kernel, return the valid cross-correlation (the operation deep-learning libraries call convolution): slide the kernel over the signal without padding and, at each position, return the sum of element-wise products. The output length is signal length minus kernel length plus 1.
Examples
Input:
[1,2,3,4], [1,0]Output:
[1,2,3]Explanation:
The kernel slides across the signal one step at a time, and each output is the dot product of the kernel with the window it currently covers.
Input:
[1,2,3], [1,1]Output:
[3,5]Explanation:
The kernel slides across the signal one step at a time, and each output is the dot product of the kernel with the window it currently covers.
Input:
[1,2,3,4,5], [1,2,1]Output:
[8,12,16]Explanation:
The kernel slides across the signal one step at a time, and each output is the dot product of the kernel with the window it currently covers.
Constraints
- •
1 ≤ kernel length ≤ signal length ≤ 10⁴