Q3. Consider the following figures 3x3 convolution of depth 1 performed over a 5x5 input feature map, also of depth 1. There are nine possible 3x3 locations to extract tiles from the 5x5 feature map ( write your roll number), so this convolution produces a 3x3 output feature map. [14 marks) 2. 2 6 Convolutional Filter 0 5 1 o o 1 1 1 1 0 7 2 0 0 1 0 Figure 2: 3x3 convolution (depth 1). Figure 1: 5x5 input feature map (depth 1) (a) Complete first Convolutional layer after apply convolutional filter on input feature map? I (b) Complete the pooling layer after applying convolutional layer (on a part) Note: Window Size of the max-pooling filter (typically 2x2 pixels) Stride: is 2