Pool.m

function pooledFeatures = cnnPool(poolDim, convolvedFeatures)
%cnnPool Pools the given convolved features
%
% Parameters:
% poolDim - dimension of pooling region
% convolvedFeatures - convolved features to pool (as given by cnnConvolve)
% convolvedFeatures(featureNum, imageNum, imageRow, imageCol)
%
% Returns:
% pooledFeatures - matrix of pooled features in the form
% pooledFeatures(featureNum, imageNum, poolRow, poolCol)
%

numImages = size(convolvedFeatures, 2);
numFeatures = size(convolvedFeatures, 1);
convolvedDim = size(convolvedFeatures, 3);
resDim = floor(convolvedDim / poolDim);

pooledFeatures = zeros(numFeatures, numImages, floor(convolvedDim / poolDim), floor(convolvedDim / poolDim));

% -------------------- YOUR CODE HERE --------------------
% Instructions:
% Now pool the convolved features in regions of poolDim x poolDim,
% to obtain the
% numFeatures x numImages x (convolvedDim/poolDim) x (convolvedDim/poolDim)
% matrix pooledFeatures, such that
% pooledFeatures(featureNum, imageNum, poolRow, poolCol) is the
% value of the featureNum feature for the imageNum image pooled over the
% corresponding (poolRow, poolCol) pooling region
% (see http://ufldl/wiki/index.php/Pooling )
%
% Use mean pooling here.
% -------------------- YOUR CODE HERE --------------------

for imageNum = 1:numImages
    for featureNum = 1:numFeatures
        for poolRow = 1:resDim
            rowStart = (poolRow - 1) * poolDim + 1;
            rowEnd = rowStart + poolDim - 1;
            for poolCol = 1:resDim
                colStart = (poolCol - 1) * poolDim + 1;
                colEnd = colStart + poolDim - 1;
                patch = convolvedFeatures(featureNum, imageNum, ...
                                          rowStart:rowEnd, ...
                                          colStart:colEnd);
                pooledFeatures(featureNum, imageNum, poolRow, poolCol) ...
                    = max(patch(:));
            end
        end
    end
end

end