If image size is 32x32 (no third dimension) and filter size is 1x1 for convolution. Then to my understanding, resulting size must be 32x32. However, if Image has third dimension like 32x32x64 and image filter is just one filter with dimensions 1x1x64, then resulting size would be 32x32 (no third dimension) and thats how the spatial dimensions would reduce. Please validate.
Asked
Active
Viewed 9 times
