Doubt at 11:06 of the video

[ishabehera (Komal Behera)]

when there are 3 conv layers, how is the 3rd layer capturing a region of 5,5 of 1st layer ?