From 5f021cb0354327f9d1c26c01231a4d7e92543731 Mon Sep 17 00:00:00 2001 From: Vasil Zlatanov Date: Tue, 20 Nov 2018 19:17:50 +0000 Subject: Change description of altenative method --- report/paper.md | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/report/paper.md b/report/paper.md index 531322e..e94e3a0 100755 --- a/report/paper.md +++ b/report/paper.md @@ -162,10 +162,10 @@ K=1, as visible in figure \ref{fig:k-diff}. \end{center} \end{figure} -The process for alternative method is somewhat similar to LDA. One different -subspace is generated for each class. These subspaces are then used for reconstruction -of the test image and the class of the subspace that generated the minimum reconstruction -error is assigned. +The process for alternative method is draws similarities to LDA. Similarly to LDA it calculates per class means. It then projects +images onto eigenvectors from subspaces generated per each class. While it does not attempt to discriminate features per class, the +calculation of independent class subspaces is effective at differentiating between the classes when reconstruction error from each class +subspace is compared. The class with the subspace that generates the least error is selected as the label. The alternative method shows overall a better performance (see figure \ref{fig:altacc}), with peak accuracy of 69% for $M=5$. The maximum $M$ non zero eigenvectors that can be used will in this case be at most -- cgit v1.2.3-54-g00ecf